- Email: [email protected]
Improved design of two-stage filter cartridges for high sulfur natural gas purification
Accepted Manuscript Improved Design of Two-stage Filter Cartridges for High Sulfur Natural Gas Purification Zhen Liu, Zhongli Ji, Jianfeng Shang, Honghai Chen, Yufeng Liu, Runpeng Wang PII: DOI: Reference:
S1383-5866(16)31660-4 http://dx.doi.org/10.1016/j.seppur.2017.05.038 SEPPUR 13751
To appear in:
Separation and Purification Technology
Received Date: Revised Date: Accepted Date:
8 September 2016 7 May 2017 22 May 2017
Please cite this article as: Z. Liu, Z. Ji, J. Shang, H. Chen, Y. Liu, R. Wang, Improved Design of Two-stage Filter Cartridges for High Sulfur Natural Gas Purification, Separation and Purification Technology (2017), doi: http:// dx.doi.org/10.1016/j.seppur.2017.05.038
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Improved Design of Two-stage Filter Cartridges for High Sulfur Natural Gas Purification Zhen Liu, Zhongli Ji*, Jianfeng Shang, Honghai Chen, Yufeng Liu, Runpeng Wang Beijing Key Laboratory of Process Fluid Filtration and Separation, College of Mechanical and Transportation Engineering, China University of Petroleum, Beijing 102249, China * Corresponding author. Tel./fax: +86 10 89734336. Address: College of Mechanical and Transportation Engineering, China University of Petroleum, Beijing 102249, China. E-mail address: [email protected].
Abstract Two-stage filtration units including pre-filtration and coalescence filtration play an important role in the quality control of high-sulfur natural gas, because micron-sized particles can significantly influence the gas sweetening process. Filtration failure usually occurs in operation, because of the mismatch between the two types of filter cartridge in the filtration unit. To overcome this drawback, the performance of the coalescing filter cartridge must be improved and a matching pre-filter cartridge need to be designed. One approach for accomplishing this is to spray an atomizing adhesive onto the fibrous layers of the filter cartridge to enhance its integrated filtration performance. In this study, a pleated protective layer and a hydrophobic drainage layer were designed for the coalescing filter cartridge to prolong the operation cycle and quickly drain the captured liquid. Testing results showed that the new-type of coalescing filter cartridge exhibited 75% decrease in its steady pressure drop over that of the original filter and the filtration efficiency reached 99.99% for liquid droplets with diameter greater than 0.3 micron. The structure of pre-filter cartridge was also redesigned to match the performance of the new coalescing filter cartridge. As a result, 1
the operational loading of the improved two-stage filter cartridges achieved the desired performance. Field testing of the new filter combination showed that the acid gas filtration unit produced a more reasonable pressure drop and a higher rate of captured liquid than with the conventional filter. Keywords: Natural Gas; Coalescing filter; Multistage Process; Filtration Performance
1. Introduction High sulfur natural gas produced from natural gas wells is transmitted to a gas purification plant through a gas gathering system. Entrained contaminants including solid particles and oil droplets are moved along with the gas to the purification plant [1]. The most common contaminants include deposits of elemental sulfur, iron sulfides/oxides, waxes, water, chemical additives, silica and sand, etc. [2] If the contaminants are not removed from the gas before reaching the desulfurization unit, the desulfurizer can become contaminated, which will result in amine solvent degradation and foaming that will affect the quality of purified gas [3]. Researchers have suggested a number of control measures to reduce amine solvent degradation and foaming, including improving the desulfurizers, adding defoamers, solution filtration and prevention of equipment corrosion, but removal of the contaminants from the feed gas is considered to be the primary measure to ameliorate [4-8]. In addition, the liquid contaminants can generate corrosive substances, and chloride in the liquid can cause stress corrosion of austenitic stainless steel [9, 10]. Therefore, a feed gas filtration unit has been established to remove the contaminants at the first step in the gas purification process, which plays a key role in the safe operation of the natural gas purification plant. Natural gas is flammable and combustible, its pressure is generally more than 4 MPa, 2
and can reach 12 MPa in practice. For this reason, natural gas filtration should be initially studied using experimental work and numerical simulations, which can be followed by safe field tests. Li et al. [11] compared the repeatability and reliability of light-scattering spectrometry and membrane filter sampling system to measure the mass concentration of outlet liquid droplets, and evaluated the gas-liquid separation performance of three types of cartridge filters under laboratory conditions. Innocentini et al. [12] tested the penetration characteristics of filter materials such as cellulose, polyester, polypropylene and stainless steel fiber, at the absolute pressures of 93 ~ 693 kPa, and found that the pressure drop across these materials increased with the pressure of the gas. Azadi et al. [13, 14] measured particle concentration and particle diameter distribution using isokinetic sampling methods at various nodes in a natural gas gathering system and concluded that the performance of the filtering system required further improvement. High sulfur natural gas contains significant quantities of hydrogen sulfide, fluctuating working conditions and a high liquid content. These conditions pose significant challenges to the conventional filtration system. Engel et al. [2] comprehensively analyzed the existing problems of filtration and separation systems in gas processing operations. These studies considered unsuitable technologies, inept compatibility, deficient vessel design, inappropriate sealing surfaces, poor media, lack of or inappropriate maintenance procedures and instrumentation deficiencies. The authors found that these factors directly produced high operation costs and/or an inability to attain sweetening specifications. Richard et al. [15] upgraded the amine filtration system in a gas plant, with amine-solution filters and an inlet gas liquid coalescing filter to achieve the projected performance of the formulated amine. In the study reported herein, the problem of filtering contaminants from natural gas
3
was analyzed in a gas purification plant. The structure of the two–stage filter cartridges was redesigned to improve the filtration performance. In addition, a matching filtration performance was conducted using a laboratory-scale evaluation system. The resulting filter cartridges were used in a field evaluation to determine the performance of the improved two-stage filtration unit. 2. Analysis of the problems in a feed gas filtration unit The feed gas filtration unit is positioned at the front of a natural gas purification system and generally contains two-stage filters, which includes two pre-filters and a coalescing filter. The feed gas filtration unit process is shown in Fig. 1. The pre-filter is used to remove particles from the gas stream that have a diameter greater than 1 micron, and the coalescing filter removes liquid droplets with diameter of 0.3 micron and more. The designed gas flow rate of the filtration unit is 300×104 Nm3/d, with the inlet pressure of 8 MPa and inlet temperature of 40 °C. The pressure drop of the filter is the parameter of interest for determining when to replace dirty filter cartridges, which is commonly set at 0.1 MPa. The pressure drop of the coalescing filter should rise with the filtration run time, since the feed gas is relatively dirty under normal operating conditions. However, it is difficult to judge when the coalescing filter cartridges required replacement based on the pressure drop of the coalescing filter, because, in practice, the pressure drop does not normally increase, even with extended use of the filter cartridge. Meanwhile, the desulfurization unit at the downstream of the filtration unit often exists amine solvent degradation and foaming, indicating that the small droplets through the filter have affected the desulfurization process. Thus, the liquid level is used to determine when to replace the filter cartridge based on the increase in the quantity of the captured liquid, which indicates that the coalescing filter is functioning properly. This poses a dilemma, 4
because it is profligate to replace filter cartridges before they have reached the end of their service life. However, if the filtration performance of the coalescence filter is less than that required for actual conditions, it is easy for particles to penetrate the filter and escape into the gas stream. The safety of the follow-up process may be adversely affected if a failed filter cartridge remains in service. By contrast, the pressure drop of the pre-filter increases too quickly, leading to the unnecessary and frequent replacement of the pre-filter cartridge. Therefore, the bases for these phenomena require close analysis. As shown in Fig. 2, a filtration performance evaluation system for the filter cartridge has been established in the laboratory, and its reliability had been verified by the previously reported research [16, 17]. In this filtration process, di-ethyl-hexyl-sebacate (DEHS) was used as the experimental liquid to generate atomized liquid aerosols. To test the performance of the two-stage filter cartridges, a pre-filter unit was placed in front of the coalescing unit. Based on the process of the feed gas filtration unit, the aerosol flows from the external potion to the internal portion of the pre-filter cartridge, and then flows from the internal to the external portion of the coalescing filter cartridge. Isokinetic sampling tubes were installed at three positions, including upstream, midstream and downstream, to produce the samples of analysis of the gas using an optical aerosol spectrometer (Welas 3000, PALAS Gmbh, Germany). This allowed for online analysis of droplet parameters including aerosol concentration and particle size distribution. Differential pressure transmitters (3051 CD, Rosemount Inc., USA) were used to record the evolution of pressure drop in the two-stage filter cartridges. The operating parameters used for these tests included a filtration velocity of 0.1 m·s-1, and an inlet aerosol concentration of 200 mg·m-3. The basis for the filter problems were
5
determined by analyzing the matching relative performance of the pre-filter and the coalescing filter. The experimental results indicated that there were two reasons for the mismatch of these two filters in field applications. One problem was based on the fact that only the pre-filter was normally active in the two-stage filtration unit. As a result, the coalescing filter is incapable of solely removing sub-micron particles. The other problem was that the two-stage filtration unit was designed without consideration of the corresponding performance of the pre-filter and coalescing filter. The main reason for the failure of the coalescing filter was considered to be the unloading of the contaminants that resulted from by the deformation of the micro-structure of the fibrous medium and the increase in porosity that resulted from fracture of the fibers in the cartridge. The scanning electron microscopic analysis of fibrous medium is shown in Fig. 3. The pressure drop of the filter increased with time during operation, and the pressure drop between the inner and outer surfaces of the filter material also gradually increased. This caused the layers of the fibrous filter material to become extruded and degraded, resulting in a variation in the porosity of the fibrous medium that in severe cases resulted in the fracture of the fiber material. The pore structure of the fiber cartridge used in the field was not stabilized by a pressure drop on both sides of the filter, it varied with the gas velocity, pressure fluctuation, mechanical vibration, and pressure increase. Under these conditions, contaminants originally attached on the surface of the fibrous filter material, can break free and become entrained in the downstream flow, causing a rise in the outlet aerosol concentration and a decrease in filtration efficiency. Also, the acid gas and liquid in the gas can corrode the fibrous medium and the metal skeleton
6
of the filter. Two resolutions were considered to improve the performance of the two-stage filtration unit. First, the structure of the coalescing filter cartridge was improved to produce low resistance and high filtration efficiency of the sub-micron droplets. Second, a matching pre-filter cartridge was designed to affect a reasonable division of labor and cooperation in the two-stage filters to lower the total pressure drop. The original pre-filter cartridge and coalescing filter cartridge were designated A1 and C1, and the improved pre-filter cartridge and coalescing filter cartridge were termed A2 and C2. 3. Improved filter cartridges design Considering the ISO 12500 standard, it is more meaningful to compare the different coalescence filters at the steady coalescence state of gas-liquid filtration process, when the filter cartridge is saturated with liquid. The laboratory conditions used for this analysis included simple droplets where the upper limit of the droplet size is 5 ~ 8 μm, and an inlet concentration of about 200 mg·m-3 to achieve a steady coalescence stage over a short period of time. This inlet concentration was much higher than that normally found under field conditions. It seemed that the liquid accumulated in the filter layers of the coalescence filter cartridge C1 could not discharge quickly, which resulted in a higher pressure drop in the steady coalescence state. The liquid outside of the filter cartridge formed a large number of bubbles in the liquid film from the action of the gas drag force, and the bubbles gradually grew and burst into many smaller droplets on the outlet side of the filter cartridge, then the captured liquid was re-entrained in the gas flow. As a result, the coalescing fibrous layers that were composed of a smaller diameter fiber, a solidification structure and hydrophobic drainage layers were selected as C2. The structure of the improved coalescing filter cartridge C2 is shown in Fig. 4. To 7
maintain the basic dimensions of the coalescing filter cartridge, a protective layer was designed to intercept solid particles thus prolonging the operation cycle. Using fibers with small diameter for the coalescence layers could improve the collection efficiency on sub-micron droplets. Adhesives were used to reinforce the multilayer fibrous layers, to improve the ability of the filter to cope with complex airflow and corrosive impurities. Also, a drainage layer composed of a hydrophobic material was placed outside of the outer frame to rapidly drain the accumulated liquid. 3.1 Pleated protective layer High sulfur gas exists within the three-phases of the gas-liquid-solid states when the gas is transported from the gathering system to the purification plant, particularly in the case of a batch process employing a corrosion inhibitor. The batch process includes the addition of a suitable quantity of corrosion inhibitors between two pigging balls, using the high pressure gas to keep the pigs moving. The metallophilic chemical bonding in the corrosion inhibitor combines with the metal surface of the pigs in the pipeline, forming a layer of protective film to protect the pipeline from corrosion [18]. Therefore, there is a large amount of particulate impurities carried with the gas to the filtration unit during batch processing. Some sub-micron particles move through the pre-filter and enter the coalescing filter, although the pre-filter cartridge can normally intercept most of the contaminants. The sub-micron particles clog the coalescence layers, which diminishes the coalescence effect. The pressure drop across the coalescing filter increases during batch processing, which shortens the operating life of the filter cartridge. Therefore, it is necessary to place a protective layer in front of the coalescence layer, to prevent the fine dust from clogging the high-precision coalescence layer and adversely affecting the droplet coalescing process. In addition, the pleated protective
8
layer can extend the service life of the coalescence filter cartridge, because of the surface filtration area of the pleated protective layer, which is five times larger than the non-pleated protective layer. [19]. Fig. 5 shows the evolution of the pressure drop during the dust filtration process. The A2 fine test dust based on the ISO 12103-1 standard was used in this reported experiment. The results show that the pressure drop of the pleated protective layer increased slowly when the same quantity of dust was collected. This result occurred, because the thickness of the cake formed on the surface of the pleated protective layer was thinner than that collected on the non-pleated protective layer, which produced a lower pressure drop. 3.2 Reinforced fibrous layer A multilayer composition of the fibrous layers is commonly used in the traditional coalescing filter, and the strength of the fiber is adapted to the operating condition. However, the linking strength between the layers is usually ignored, since the layers normally deform at high pressure, high sulfur content or fluctuations in the gas flow. To address this problem, the use of an adhesive was proposed to position the layers. The adhesive was applied to the surface fibrous layer of the filter using micron-sized colloidal particles that were generated by an atomized spray nozzle employing compressed air. The strength of each layer was enhanced by the adhesive and the integrity of the web structure was ensured. As a result, the captured particles were not released in the downstream gas flow, which preserved the integrity of the filtration process. After application of the adhesive, the various layers were sequentially overlaid and the cured adhesive bonded the layers together. This multi-layer fibrous structure will not change under fluctuations in gas velocity or pressure, mechanical vibration or even as a result of an increase in pressure drop. In
9
addition, this method can avoid fiber shedding and increase the effective filtration area for removing particulate contaminants, which will indirectly improve filtration performance. A waterborne aliphatic polyurethane was selected from a large number of possible candidates for use as the adhesive in this study. The results of the adhesive spraying technique are shown in Fig. 6, where the yellow circles represent the presence of the adhesive on the fibrous net. Two types of filter material with and without adhesive were selected to test the filter tear strength, the test filter was a circular piece of material with an area of 0.01 m2. Three samples of each type of filter material were tested using a universal material testing machine. The tests complied with the Standard ISO 13934-1-2013 for textiles (Tensile properties of fabrics - Part 1: Determination of maximum force and elongation at maximum force using the strip method). The test results showed that the average maximum tearing strength values for the filter material containing the adhesive was 21.3 N and the shear value for the filter material without the adhesive was 13.7 N. The fiber strength of filter material containing the adhesive was about 55% higher than the original material, which illustrated the feasibility of preserving the fiber structure by applying an adhesive to the filter material. The working conditions in a field site were simulated in the laboratory to test the acid resistance of the adhesive. A sample of the filter material containing the adhesive was dipped in sulphuric acid with the concentration of 0.025 mol/L for three days, and then the treated filter material was then subjected to the tear test. The tear test complied with the Chinese National Standard GB/T 17632-1998 for textiles (Geotextiles and geotextile-related products - test method for determining the resistance to liquids of acids and bases). The test results showed that the average breaking strength of filter material before acid immersion was 24.7 N and after acid
10
immersion the material exhibited a tear strength of 24.1 N. Since 97.6% of the tear strength was maintained after acid immersion, it was concluded that the adhesive provided acceptable strength to the material after exposure to a strong acid. It should be noted that the adhesive appeared to attach to the surface of fiber material, which could decrease the porosity and effective area for gas flow. Therefore, a comparative experiment was conducted using both the untreated and reinforced filter layers, to determine the effect of the adhesive on the filtration performance, including the initial pressure drop, the pressure drop during filtration and the filtration efficiency. Fig. 7 shows the initial pressure drop of the material at various filtration gas velocities. As shown, the values for the reinforced filter layers increased by 23% over those of the untreated filter layers, which is an acceptable result. The pressure drops across the untreated and reinforced filter layers are shown in Fig. 8. It can be seen that the porosity of reinforced filter layers decreased, because of the adhesive, which decreased the period of liquid filling within the filter layers. As shown, the reinforced filter layers achieved a steady state faster than the untreated filter layers. The evolution of the pressure drop of untreated filter layers followed the usual trend for liquid aerosol filtration, remaining constant after an initial period of rapid increase [20, 21]. However, an unexpected phenomenon occurred during this process where the pressure drop of the reinforced filter layers decreased slowly toward the achievement of a steady state. Also, the liquid that was discharged from the reinforced filter layers appeared to be slightly turbid, whereas the liquid of the untreated filter cartridge was clear. Based on the observation and analysis, it was concluded that a variety of additives contained in the adhesive that were intended to improve bonding, did not attach to the fibers after the adhesive cured, and leached out
11
of the fibers into the entrained liquid. Fig. 9 shows the accumulated filtration efficiency of the untreated and reinforced filter layers. Normally, the overall filtration efficiency of the filter layers containing adhesives should increase as a result of the increased packing density and compact fiber web. However, the efficiency of the reinforced filter layer was slightly lower when the droplets were larger than 0.8 microns. Under the identical inlet concentration conditions, the treated filter material had greater number of droplets with a diameter of more than 0.8 microns than the untreated filter layer. It is believed that the liquid viscosity, surface tension and other physical properties were changed as a result of the mixing of the adhesive additives with the captured liquid droplets. This complicated the drainage process, and resulted in more re-entrainment of droplets. This conclusion can be verified by observations during the experimental process. The color of the liquid obtained from the reinforced filter layers was more turbid, and bubbles were generated at the outlet side of the layers. In addition, this phenomenon would be improved through placing a hydrophobic drainage layer after the reinforced fibrous layer. In general, application of adhesives to the filter appeared to enhance the structural strength of the filter material, and caused no obvious negative effects to the filtration efficiency. This method could be used for the industrial manufacture of filter cartridges after optimizing the processing technology. 3.3 Hydrophobic drainage layer A gas-liquid filtration process includes a complex process for liquid droplet interception, coalescence, migration and entrainment. The liquid outside the filter cartridge would form a large number of bubbles from the liquid film under the action of the gas drag force, and the bubbles gradually grow and burst into lot of droplets at the outlet side of the filter cartridge, then the captured liquid was re-entrained with gas flow.
12
Related research of liquid aerosol filtration by hydrophobic filter materials, has shown that hydrophobic filter material can significantly reduce liquid saturation and the risk of liquid re-entrainment in the gas stream [22-24]. Therefore, to further improve the properties of the subject filters, a layer of hydrophobic polyester fiber was placed in the filter after the coalescence layer and transition layer. A comparative test of the gas-liquid filtration performance was then conducted under the same experimental conditions previously employed, to determine the effect of the addition of this hydrophobic layer. The results of the tests are shown in Fig. 10. As shown, it appeared that the new-type of coalescing filter cartridge reached the steady state earlier, and the pressure drop of C2 was only 25% of C1 under the same operating conditions. The 0.3 micron droplet filtration efficiency reached 99.99%, indicating that the filtration performance of C2 was significantly improved. 3.4 Performance matching of the two-stage filter cartridges The degradation of filtration performance of the pre-filter cartridge was reduced by adjusting the structure and parameters of filter medium. More specifically, the design of a new type pre-filter cartridge was based on three aspects as shown in Fig. 11. The pleated layer with a high dust holding capacity was designed as the first layer at the inlet side to cope with the larger contaminant particles, with special attention to the operating conditions of pigging and batch processing. A group of fine filter layers was placed in the cartridge to capture the fine particles that penetrated the first layer, to ensure filtration precision. A frame was set at the inlet side to ensure that the filter would not be crushed when the pressure drop was large and the cartridge frame material was 316L stainless steel to avoid the corrosion in acid environments. Eventually, the pressure drop across A2 decreased by 50% over that of A1. This was because by the increase in droplets with a diameter of less than 2 micron that
13
penetrated through A2 and were captured by the coalescing filter. Therefore, the operating load of the two-stage filters received a reasonable allocation of droplets. Fig. 11 shows the pressure drop for the combination of A2 and C2. The pressure drop in both A2 and C2 gradually increased and the total pressure drop decreased. The droplets with a diameter of more than 2 micron were nearly all removed from the gas stream after A2 and the outlet concentration of droplets from C2 was relatively low. However, the practical effects of the improved filter cartridges remained to be demonstrated in actual field application. 4. Application of improved filter cartridges A combination unit in a natural gas purification plant in southwest China was chosen as the site to conduct the field tests on the improved filter cartridges. This plant had two series units as a combination unit. The conventional filter cartridges were separately installed as the control group and the improved filter cartridges were installed as the trial group. Process data were recorded during the test, including the inlet gas pressure, the gas flow from the filtration unit, the pressure drop and liquid drainage volume of the pre-filter and coalescing filter. The data were accumulated for 800 hours of testing and were selected for a comparative analysis of the filters. The inlet pressure was stable at about 8 MPa during the field test and the gas flows of the two series were nearly the identical. For the trial group, the pressure drop of the two-stage filters have risen to 15 kPa and 4 kPa, while the values for the control group showed no obvious change. The increased pressure drop was due primarily to the accumulation of solid particles that clogged the filter layers. Therefore, the increase in the pressure drop between the two groups showed that the trial group filter exhibited better filtration performance. During the 800 hours of testing, the liquid that accumulated in the control group filters 14
was drained 15 times for a cumulative drainage volume of 0.95 m3. By contrast, the trial group filters were drained 28 times with a drainage volume of 2.03 m3. These results indicated that the trial group exhibited more than twice the liquid intercept capability of the control group. These results showed that the improved filter combination had significantly improved filtration performance for both solid particles and liquid droplets. Also, the two-stage filter exhibited a synchronous increase in filtration, which achieved the goal of performance optimization and matching of the two-stage filtration unit in a natural gas purification plant. 5. Conclusions The feed gas in a natural gas purification plant contains high concentrations of hydrogen sulfide, high liquid content and experiences fluctuating working conditions. Factors including operating conditions, filter performance and operation costs, must be comprehensively considered when choosing the type and materials of the filter cartridges used to clean up the gas stream. This reported work focused on the problems existing in natural gas filtration plants where the pressure drop of a coalescing filter and a pre-filter are mismatched. Comprehensive analysis of the problem showed that in the two-stage filtration unit that was designed without matching the performance of the pre-filter and coalescing filter, the coalescing filter was individually incapable of removing sub-micron particles. Methods were presented for improving the filtration efficiency and stability of coalescing filter cartridge. To obtain low resistance and highly efficient operation in a two-stage filtration unit, reasonable cooperation and performance matching were examined, together with optimization of the performance of the pre-filter cartridge and the coalescing filter cartridge. Based on the experimental results of these studies, especially the method of improving the coalescing filter cartridge and performance matching of two-stage filter 15
cartridges, this present study provided a reference for designing high-pressure and corrosive gas filtration system. Acknowledgement This work was supported by China National Natural Science Foundation (No. 51376196) and China National Science and Technology Major Project (No. 2016ZX05017). References [1]
Y. Alcheikhhamdon, M. Hoorfar, Natural gas quality enhancement: A review of the conventional treatment processes, and the industrial challenges facing emerging technologies, J. Nat. Gas Sci. Eng., 34 (2016) 689-701.
[2]
D. Engel, M.H. Sheilan, Seven Deadly Sins of Filtration and Separation Systems in Gas Processing Operations, in: Y. Wu, J.J. Carroll, Q. Li (Eds.) Gas Injection for Disposal and Enhanced Recovery, John Wiley & Sons, Inc., Beverly, 2014, pp. 227-241.
[3]
P. Cézac, J.P. Serin, J.M. Reneaume, J. Mercadier, G. Mouton, Elemental sulphur deposition in natural gas transmission and distribution networks, J. Supercrit. Fluid., 44 (2008) 115-122.
[4]
E. Alhseinat, P. Pal, A. Ganesan, F. Banat, Effect of MDEA degradation products on foaming behavior and physical properties of aqueous MDEA solutions, Int. J. Greenh. Gas Con., 37 (2015) 280-286.
[5]
P. Pal, A. AbuKashabeh, S. Al-Asheh, F. Banat, Role of aqueous methyldiethanolamine (MDEA) as solvent in natural gas sweetening unit and process contaminants with probable reaction pathway, J. Nat. Gas Sci. Eng., 24 (2015) 124-131.
[6]
Y. Ke, B. Shen, H. Sun, J. Liu, X. Xu, Study on foaming of formulated solvent UDS and improving foaming control in acid natural gas sweetening process, J. Nat. Gas Sci. Eng., 28 (2016) 271-279.
[7]
C.R. Pauley, R. Hashemi, S. Caothien, C.R. Pauley, S. Caothien, Ways to control amine unit foaming offered, Oil Gas J., 87 (1989) 67-75.
[8]
T. Hu, G. Xiong, J. He, J. Yin, X. Peng, Prevention and control measures for 16
foaming in amine desulfurization unit, Natural Gas Industry, 29 (2009) 101-103. [9]
Z. Wang, C. Yang, L. Zhu, Corrosion analysis and inhibition studies in the process of natural gas wet desulfurization, Eng. Fail. Anal. , 44 (2014) 66-73.
[10]
F. Li, G. Sun, Q. Zhang, J. Long, Corrosion behavior and prevention of gas sweetening unit in natural gas processing plants, Natural Gas Industry, (2009) 104-106.
[11]
B. Li, Z. Ji, X. Yang, Evaluation of gas-liquid separation performance of natural gas filters, Pet. Sci., 6 (2009) 438-444.
[12]
M.D.M. Innocentini, E.H. Tanabe, M.L. Aguiar, J.R. Coury, Filtration of gases at high pressures: Permeation behavior of fiber-based media used for natural gas cleaning, Chem. Eng. Sci., 74 (2012) 38-48.
[13]
M. Azadi, A. Mohebbi, S. Soltaninejad, F. Scala, A case study on suspended particles in a natural gas urban transmission and distribution network, Fuel Process Technol., 93 (2012) 65-72.
[14]
M. Azadi, A. Mohebbi, F. Scala, S. Soltaninejad, Experimental study of filtration system performance of natural gas in urban transmission and distribution network: A case study on the city of Kerman, Iran, Fuel, 90 (2011) 1166-1171.
[15]
C.R. Pauley, D.G. Langston, F.C. Betts, Redesigned filters solve foaming, amine-loss problems at Louisiana gas plant, Oil Gas J., 89 (1991) 50-55.
[16]
Z. Liu, Z. Ji, X. Wu, H. Ma, F. Zhao, Y. Hao, Experimental investigation on liquid distribution of filter cartridge during gas-liquid filtration, Sep. Purif. Technol., 170 (2016) 146-154.
[17]
C. Chang, Z. Ji, C. Liu, F. Zhao, Permeability of filter cartridges used for natural gas filtration at high pressure, J. Nat. Gas Sci. Eng., 34 (2016) 419-427.
[18]
M. Finšgar, J. Jackson, Application of corrosion inhibitors for steels in acidic media for the oil and gas industry: A review, Corros. Sci., 86 (2014) 17-41.
[19]
A.M. Saleh, H.V. Tafreshi, B. Pourdeyhimi, Service life of circular pleated filters vs. that of their flat counterpart, Sep. Purif. Technol., 156, Part 2 (2015) 881-888.
[20]
A. Charvet, Y. Gonthier, E. Gonze, A. Bernis, Experimental and modelled efficiencies during the filtration of a liquid aerosol with a fibrous medium, Chem. Eng. Sci., 65 (2010) 1875-1886. 17
[21]
P. Contal, J. Simao, D. Thomas, T. Frising, S. Callé, J.C. Appert-Collin, D. Bémer, Clogging of fibre filters by submicron droplets. Phenomena and influence of operating conditions, J. Aerosol Sci., 35 (2004) 263-278.
[22]
B.J. Mullins, R. Mead Hunter, R.N. Pitta, G. Kasper, W. Heikamp, Comparative performance of philic and phobic oil‐ mist filters, AIChE J., (2014).
[23]
C. Chang, Z. Ji, F. Zeng, The effect of a drainage layer on filtration performance of coalescing filters, Sep. Purif. Technol., 170 (2016) 370-376.
[24]
R. Mead-Hunter, A.J.C. King, B.J. Mullins, Aerosol-mist coalescing filters - A review, Sep. Purif. Technol., 133 (2014) 484-506.
18
Figure Legends: Fig. 1 Feed gas filtration unit in natural gas purification system. Fig. 2 Filtration performance evaluation system for two-stage filter cartridges. Fig. 3 Main reason for failure of the fibrous medium. (the schematic diagram was excerpted from the literature published by Pall Corporation, USA) Fig. 4 Improved coalescing filter cartridge design. Fig. 5 Influence of the protecting layer on the pressure drop of coalescing filter cartridge C1. Fig. 6 SEM micrographs of fiber net with/without adhesive. Fig. 7 Influence of adhesive on initial pressure drop of filter layers. Fig. 8 Influence of adhesive on pressure drop of filter layers during gas-liquid filtration. Fig. 9 Influence of adhesive on filtration efficiency of filter layers. Fig. 10 Comparison of gas-liquid filtration performance of coalescing filter cartridges. Fig. 11 Structure design of pre-filter cartridge. Fig. 12 Pressure drop evolution for the combination of A2 and C2.
19
ΔP
Clean gas Feed gas
Pre-filter A
ΔP
Coalescer C ΔP
Standby Pre-filter B
Fig. 13 Feed gas filtration unit in natural gas purification system.
20
Compressed air
Upstream sampling
Downstream sampling
Mixing chamber
Liquid
Coalescing filter
Pre-filter
Protect filter PID control
Spraying nozzle Aerosol particle size spectrometer
PID control
Flowmeter Valve V-3
T&H
Air Conditioner
Liquid Gatherer
Air
Blower
ΔP
ΔP
Midstream sampling
Inlet filter
Liquid Gatherer
Fig. 14 Filtration performance evaluation system for two-stage filter cartridges.
21
a. Contaminant unloading
b. Structure damage
104μm
573μm
Fig. 15 Main reason for failure of the fibrous medium. (the schematic diagram was excerpted from the literature published by Pall Corporation, USA)
22
Coalescence layer Transition layer Adhesive particles Outer frame Protective layer
Inner frame
Drainage layer
Fig. 16 Improved coalescing filter cartridge design.
23
Original Coalecing filter C1 C1 with pleated protecting layer
Pressure drop (kPa)
2.0
1.5
1.0
0.5
0.0
0
20 40 2 Dust loading per unit filter area (g/m )
60
Fig. 17 Influence of the protecting layer on the pressure drop of coalescing filter cartridge C1.
24
a. Clean state
b. With adhesive
Fig. 18 SEM micrographs of fiber net with/without adhesive.
25
1.0
Pressure drop (kPa)
0.8
0.6
0.4
Untreated filter layers Reinforced filter layers
0.2
0.0 0.04
0.06
0.08
0.10
0.12
0.14
Face velocity (m/s)
Fig. 19 Influence of adhesive on initial pressure drop of filter layers.
26
4.0
Pressure drop（ kPa）
3.5 3.0 2.5 2.0 1.5 1.0
Untreated filter layers Reinforced filter layers
0.5 0.0
0
1
2
3
4
5
6
Time (h)
Fig. 20 Influence of adhesive on pressure drop of filter layers during gas-liquid filtration.
27
Filtration efficiency (%)
100.0 99.9 99.8 99.7 99.6
Untreated filter layers Reinforced filter layers
99.5 99.4
0.3
0.5
1 2 Droplet size (μm)
4
8
Fig. 21 Influence of adhesive on filtration efficiency of filter layers.
28
16
8 kPa
Pressure drop (kPa)
8 4
2 kPa
2 1 0.5
Coalescing filter cartridge C1 Coalescing filter cartridge C2
0.25 0
1
2
3
4
5
Time (h)
a. Pressure drop evolution
Cumulative efficiency (%)
99.9999 99.999
99.99%
99.99
99.9%
99.9 99.5 99 98
Coalescing filter cartridge C2 Coalescing filter cartridge C1
95 0.3
0.5
1
2
4
8
Droplet size (¦ Ìm)
b. Filtration efficiency at steady state Fig. 22 Comparison of gas-liquid filtration performance of coalescing filter cartridges.
29
a. Original type Single medium without frame
b. New type High dust holding layer Fine filter layer
Stainless steel inner frame
Fig. 23 Structure design of pre-filter cartridge.
30
3.0
Pre-filter cartridge A2 Coalescing filter cartridge C2
Pressure drop (kPa)
2.5 2.0 1.5 1.0 0.5 0.0
0
2
4
6
8
10
12
14
Time (h)
Fig. 24 Pressure drop evolution for the combination of A2 and C2.
31
Highlights (1) The failure of the coalescing filter mostly results from the deformation of fibrous structure and the fracture of the fibers. (2) Application of adhesives to the filter could enhance the structural strength of the filter material. (3) The two-stage filtration unit was designed with consideration of the performance matching of the pre-filter and coalescing filter.
32
S1383-5866(16)31660-4 http://dx.doi.org/10.1016/j.seppur.2017.05.038 SEPPUR 13751
To appear in:
Separation and Purification Technology
Received Date: Revised Date: Accepted Date:
8 September 2016 7 May 2017 22 May 2017
Please cite this article as: Z. Liu, Z. Ji, J. Shang, H. Chen, Y. Liu, R. Wang, Improved Design of Two-stage Filter Cartridges for High Sulfur Natural Gas Purification, Separation and Purification Technology (2017), doi: http:// dx.doi.org/10.1016/j.seppur.2017.05.038
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Improved Design of Two-stage Filter Cartridges for High Sulfur Natural Gas Purification Zhen Liu, Zhongli Ji*, Jianfeng Shang, Honghai Chen, Yufeng Liu, Runpeng Wang Beijing Key Laboratory of Process Fluid Filtration and Separation, College of Mechanical and Transportation Engineering, China University of Petroleum, Beijing 102249, China * Corresponding author. Tel./fax: +86 10 89734336. Address: College of Mechanical and Transportation Engineering, China University of Petroleum, Beijing 102249, China. E-mail address: [email protected].
Abstract Two-stage filtration units including pre-filtration and coalescence filtration play an important role in the quality control of high-sulfur natural gas, because micron-sized particles can significantly influence the gas sweetening process. Filtration failure usually occurs in operation, because of the mismatch between the two types of filter cartridge in the filtration unit. To overcome this drawback, the performance of the coalescing filter cartridge must be improved and a matching pre-filter cartridge need to be designed. One approach for accomplishing this is to spray an atomizing adhesive onto the fibrous layers of the filter cartridge to enhance its integrated filtration performance. In this study, a pleated protective layer and a hydrophobic drainage layer were designed for the coalescing filter cartridge to prolong the operation cycle and quickly drain the captured liquid. Testing results showed that the new-type of coalescing filter cartridge exhibited 75% decrease in its steady pressure drop over that of the original filter and the filtration efficiency reached 99.99% for liquid droplets with diameter greater than 0.3 micron. The structure of pre-filter cartridge was also redesigned to match the performance of the new coalescing filter cartridge. As a result, 1
the operational loading of the improved two-stage filter cartridges achieved the desired performance. Field testing of the new filter combination showed that the acid gas filtration unit produced a more reasonable pressure drop and a higher rate of captured liquid than with the conventional filter. Keywords: Natural Gas; Coalescing filter; Multistage Process; Filtration Performance
1. Introduction High sulfur natural gas produced from natural gas wells is transmitted to a gas purification plant through a gas gathering system. Entrained contaminants including solid particles and oil droplets are moved along with the gas to the purification plant [1]. The most common contaminants include deposits of elemental sulfur, iron sulfides/oxides, waxes, water, chemical additives, silica and sand, etc. [2] If the contaminants are not removed from the gas before reaching the desulfurization unit, the desulfurizer can become contaminated, which will result in amine solvent degradation and foaming that will affect the quality of purified gas [3]. Researchers have suggested a number of control measures to reduce amine solvent degradation and foaming, including improving the desulfurizers, adding defoamers, solution filtration and prevention of equipment corrosion, but removal of the contaminants from the feed gas is considered to be the primary measure to ameliorate [4-8]. In addition, the liquid contaminants can generate corrosive substances, and chloride in the liquid can cause stress corrosion of austenitic stainless steel [9, 10]. Therefore, a feed gas filtration unit has been established to remove the contaminants at the first step in the gas purification process, which plays a key role in the safe operation of the natural gas purification plant. Natural gas is flammable and combustible, its pressure is generally more than 4 MPa, 2
and can reach 12 MPa in practice. For this reason, natural gas filtration should be initially studied using experimental work and numerical simulations, which can be followed by safe field tests. Li et al. [11] compared the repeatability and reliability of light-scattering spectrometry and membrane filter sampling system to measure the mass concentration of outlet liquid droplets, and evaluated the gas-liquid separation performance of three types of cartridge filters under laboratory conditions. Innocentini et al. [12] tested the penetration characteristics of filter materials such as cellulose, polyester, polypropylene and stainless steel fiber, at the absolute pressures of 93 ~ 693 kPa, and found that the pressure drop across these materials increased with the pressure of the gas. Azadi et al. [13, 14] measured particle concentration and particle diameter distribution using isokinetic sampling methods at various nodes in a natural gas gathering system and concluded that the performance of the filtering system required further improvement. High sulfur natural gas contains significant quantities of hydrogen sulfide, fluctuating working conditions and a high liquid content. These conditions pose significant challenges to the conventional filtration system. Engel et al. [2] comprehensively analyzed the existing problems of filtration and separation systems in gas processing operations. These studies considered unsuitable technologies, inept compatibility, deficient vessel design, inappropriate sealing surfaces, poor media, lack of or inappropriate maintenance procedures and instrumentation deficiencies. The authors found that these factors directly produced high operation costs and/or an inability to attain sweetening specifications. Richard et al. [15] upgraded the amine filtration system in a gas plant, with amine-solution filters and an inlet gas liquid coalescing filter to achieve the projected performance of the formulated amine. In the study reported herein, the problem of filtering contaminants from natural gas
3
was analyzed in a gas purification plant. The structure of the two–stage filter cartridges was redesigned to improve the filtration performance. In addition, a matching filtration performance was conducted using a laboratory-scale evaluation system. The resulting filter cartridges were used in a field evaluation to determine the performance of the improved two-stage filtration unit. 2. Analysis of the problems in a feed gas filtration unit The feed gas filtration unit is positioned at the front of a natural gas purification system and generally contains two-stage filters, which includes two pre-filters and a coalescing filter. The feed gas filtration unit process is shown in Fig. 1. The pre-filter is used to remove particles from the gas stream that have a diameter greater than 1 micron, and the coalescing filter removes liquid droplets with diameter of 0.3 micron and more. The designed gas flow rate of the filtration unit is 300×104 Nm3/d, with the inlet pressure of 8 MPa and inlet temperature of 40 °C. The pressure drop of the filter is the parameter of interest for determining when to replace dirty filter cartridges, which is commonly set at 0.1 MPa. The pressure drop of the coalescing filter should rise with the filtration run time, since the feed gas is relatively dirty under normal operating conditions. However, it is difficult to judge when the coalescing filter cartridges required replacement based on the pressure drop of the coalescing filter, because, in practice, the pressure drop does not normally increase, even with extended use of the filter cartridge. Meanwhile, the desulfurization unit at the downstream of the filtration unit often exists amine solvent degradation and foaming, indicating that the small droplets through the filter have affected the desulfurization process. Thus, the liquid level is used to determine when to replace the filter cartridge based on the increase in the quantity of the captured liquid, which indicates that the coalescing filter is functioning properly. This poses a dilemma, 4
because it is profligate to replace filter cartridges before they have reached the end of their service life. However, if the filtration performance of the coalescence filter is less than that required for actual conditions, it is easy for particles to penetrate the filter and escape into the gas stream. The safety of the follow-up process may be adversely affected if a failed filter cartridge remains in service. By contrast, the pressure drop of the pre-filter increases too quickly, leading to the unnecessary and frequent replacement of the pre-filter cartridge. Therefore, the bases for these phenomena require close analysis. As shown in Fig. 2, a filtration performance evaluation system for the filter cartridge has been established in the laboratory, and its reliability had been verified by the previously reported research [16, 17]. In this filtration process, di-ethyl-hexyl-sebacate (DEHS) was used as the experimental liquid to generate atomized liquid aerosols. To test the performance of the two-stage filter cartridges, a pre-filter unit was placed in front of the coalescing unit. Based on the process of the feed gas filtration unit, the aerosol flows from the external potion to the internal portion of the pre-filter cartridge, and then flows from the internal to the external portion of the coalescing filter cartridge. Isokinetic sampling tubes were installed at three positions, including upstream, midstream and downstream, to produce the samples of analysis of the gas using an optical aerosol spectrometer (Welas 3000, PALAS Gmbh, Germany). This allowed for online analysis of droplet parameters including aerosol concentration and particle size distribution. Differential pressure transmitters (3051 CD, Rosemount Inc., USA) were used to record the evolution of pressure drop in the two-stage filter cartridges. The operating parameters used for these tests included a filtration velocity of 0.1 m·s-1, and an inlet aerosol concentration of 200 mg·m-3. The basis for the filter problems were
5
determined by analyzing the matching relative performance of the pre-filter and the coalescing filter. The experimental results indicated that there were two reasons for the mismatch of these two filters in field applications. One problem was based on the fact that only the pre-filter was normally active in the two-stage filtration unit. As a result, the coalescing filter is incapable of solely removing sub-micron particles. The other problem was that the two-stage filtration unit was designed without consideration of the corresponding performance of the pre-filter and coalescing filter. The main reason for the failure of the coalescing filter was considered to be the unloading of the contaminants that resulted from by the deformation of the micro-structure of the fibrous medium and the increase in porosity that resulted from fracture of the fibers in the cartridge. The scanning electron microscopic analysis of fibrous medium is shown in Fig. 3. The pressure drop of the filter increased with time during operation, and the pressure drop between the inner and outer surfaces of the filter material also gradually increased. This caused the layers of the fibrous filter material to become extruded and degraded, resulting in a variation in the porosity of the fibrous medium that in severe cases resulted in the fracture of the fiber material. The pore structure of the fiber cartridge used in the field was not stabilized by a pressure drop on both sides of the filter, it varied with the gas velocity, pressure fluctuation, mechanical vibration, and pressure increase. Under these conditions, contaminants originally attached on the surface of the fibrous filter material, can break free and become entrained in the downstream flow, causing a rise in the outlet aerosol concentration and a decrease in filtration efficiency. Also, the acid gas and liquid in the gas can corrode the fibrous medium and the metal skeleton
6
of the filter. Two resolutions were considered to improve the performance of the two-stage filtration unit. First, the structure of the coalescing filter cartridge was improved to produce low resistance and high filtration efficiency of the sub-micron droplets. Second, a matching pre-filter cartridge was designed to affect a reasonable division of labor and cooperation in the two-stage filters to lower the total pressure drop. The original pre-filter cartridge and coalescing filter cartridge were designated A1 and C1, and the improved pre-filter cartridge and coalescing filter cartridge were termed A2 and C2. 3. Improved filter cartridges design Considering the ISO 12500 standard, it is more meaningful to compare the different coalescence filters at the steady coalescence state of gas-liquid filtration process, when the filter cartridge is saturated with liquid. The laboratory conditions used for this analysis included simple droplets where the upper limit of the droplet size is 5 ~ 8 μm, and an inlet concentration of about 200 mg·m-3 to achieve a steady coalescence stage over a short period of time. This inlet concentration was much higher than that normally found under field conditions. It seemed that the liquid accumulated in the filter layers of the coalescence filter cartridge C1 could not discharge quickly, which resulted in a higher pressure drop in the steady coalescence state. The liquid outside of the filter cartridge formed a large number of bubbles in the liquid film from the action of the gas drag force, and the bubbles gradually grew and burst into many smaller droplets on the outlet side of the filter cartridge, then the captured liquid was re-entrained in the gas flow. As a result, the coalescing fibrous layers that were composed of a smaller diameter fiber, a solidification structure and hydrophobic drainage layers were selected as C2. The structure of the improved coalescing filter cartridge C2 is shown in Fig. 4. To 7
maintain the basic dimensions of the coalescing filter cartridge, a protective layer was designed to intercept solid particles thus prolonging the operation cycle. Using fibers with small diameter for the coalescence layers could improve the collection efficiency on sub-micron droplets. Adhesives were used to reinforce the multilayer fibrous layers, to improve the ability of the filter to cope with complex airflow and corrosive impurities. Also, a drainage layer composed of a hydrophobic material was placed outside of the outer frame to rapidly drain the accumulated liquid. 3.1 Pleated protective layer High sulfur gas exists within the three-phases of the gas-liquid-solid states when the gas is transported from the gathering system to the purification plant, particularly in the case of a batch process employing a corrosion inhibitor. The batch process includes the addition of a suitable quantity of corrosion inhibitors between two pigging balls, using the high pressure gas to keep the pigs moving. The metallophilic chemical bonding in the corrosion inhibitor combines with the metal surface of the pigs in the pipeline, forming a layer of protective film to protect the pipeline from corrosion [18]. Therefore, there is a large amount of particulate impurities carried with the gas to the filtration unit during batch processing. Some sub-micron particles move through the pre-filter and enter the coalescing filter, although the pre-filter cartridge can normally intercept most of the contaminants. The sub-micron particles clog the coalescence layers, which diminishes the coalescence effect. The pressure drop across the coalescing filter increases during batch processing, which shortens the operating life of the filter cartridge. Therefore, it is necessary to place a protective layer in front of the coalescence layer, to prevent the fine dust from clogging the high-precision coalescence layer and adversely affecting the droplet coalescing process. In addition, the pleated protective
8
layer can extend the service life of the coalescence filter cartridge, because of the surface filtration area of the pleated protective layer, which is five times larger than the non-pleated protective layer. [19]. Fig. 5 shows the evolution of the pressure drop during the dust filtration process. The A2 fine test dust based on the ISO 12103-1 standard was used in this reported experiment. The results show that the pressure drop of the pleated protective layer increased slowly when the same quantity of dust was collected. This result occurred, because the thickness of the cake formed on the surface of the pleated protective layer was thinner than that collected on the non-pleated protective layer, which produced a lower pressure drop. 3.2 Reinforced fibrous layer A multilayer composition of the fibrous layers is commonly used in the traditional coalescing filter, and the strength of the fiber is adapted to the operating condition. However, the linking strength between the layers is usually ignored, since the layers normally deform at high pressure, high sulfur content or fluctuations in the gas flow. To address this problem, the use of an adhesive was proposed to position the layers. The adhesive was applied to the surface fibrous layer of the filter using micron-sized colloidal particles that were generated by an atomized spray nozzle employing compressed air. The strength of each layer was enhanced by the adhesive and the integrity of the web structure was ensured. As a result, the captured particles were not released in the downstream gas flow, which preserved the integrity of the filtration process. After application of the adhesive, the various layers were sequentially overlaid and the cured adhesive bonded the layers together. This multi-layer fibrous structure will not change under fluctuations in gas velocity or pressure, mechanical vibration or even as a result of an increase in pressure drop. In
9
addition, this method can avoid fiber shedding and increase the effective filtration area for removing particulate contaminants, which will indirectly improve filtration performance. A waterborne aliphatic polyurethane was selected from a large number of possible candidates for use as the adhesive in this study. The results of the adhesive spraying technique are shown in Fig. 6, where the yellow circles represent the presence of the adhesive on the fibrous net. Two types of filter material with and without adhesive were selected to test the filter tear strength, the test filter was a circular piece of material with an area of 0.01 m2. Three samples of each type of filter material were tested using a universal material testing machine. The tests complied with the Standard ISO 13934-1-2013 for textiles (Tensile properties of fabrics - Part 1: Determination of maximum force and elongation at maximum force using the strip method). The test results showed that the average maximum tearing strength values for the filter material containing the adhesive was 21.3 N and the shear value for the filter material without the adhesive was 13.7 N. The fiber strength of filter material containing the adhesive was about 55% higher than the original material, which illustrated the feasibility of preserving the fiber structure by applying an adhesive to the filter material. The working conditions in a field site were simulated in the laboratory to test the acid resistance of the adhesive. A sample of the filter material containing the adhesive was dipped in sulphuric acid with the concentration of 0.025 mol/L for three days, and then the treated filter material was then subjected to the tear test. The tear test complied with the Chinese National Standard GB/T 17632-1998 for textiles (Geotextiles and geotextile-related products - test method for determining the resistance to liquids of acids and bases). The test results showed that the average breaking strength of filter material before acid immersion was 24.7 N and after acid
10
immersion the material exhibited a tear strength of 24.1 N. Since 97.6% of the tear strength was maintained after acid immersion, it was concluded that the adhesive provided acceptable strength to the material after exposure to a strong acid. It should be noted that the adhesive appeared to attach to the surface of fiber material, which could decrease the porosity and effective area for gas flow. Therefore, a comparative experiment was conducted using both the untreated and reinforced filter layers, to determine the effect of the adhesive on the filtration performance, including the initial pressure drop, the pressure drop during filtration and the filtration efficiency. Fig. 7 shows the initial pressure drop of the material at various filtration gas velocities. As shown, the values for the reinforced filter layers increased by 23% over those of the untreated filter layers, which is an acceptable result. The pressure drops across the untreated and reinforced filter layers are shown in Fig. 8. It can be seen that the porosity of reinforced filter layers decreased, because of the adhesive, which decreased the period of liquid filling within the filter layers. As shown, the reinforced filter layers achieved a steady state faster than the untreated filter layers. The evolution of the pressure drop of untreated filter layers followed the usual trend for liquid aerosol filtration, remaining constant after an initial period of rapid increase [20, 21]. However, an unexpected phenomenon occurred during this process where the pressure drop of the reinforced filter layers decreased slowly toward the achievement of a steady state. Also, the liquid that was discharged from the reinforced filter layers appeared to be slightly turbid, whereas the liquid of the untreated filter cartridge was clear. Based on the observation and analysis, it was concluded that a variety of additives contained in the adhesive that were intended to improve bonding, did not attach to the fibers after the adhesive cured, and leached out
11
of the fibers into the entrained liquid. Fig. 9 shows the accumulated filtration efficiency of the untreated and reinforced filter layers. Normally, the overall filtration efficiency of the filter layers containing adhesives should increase as a result of the increased packing density and compact fiber web. However, the efficiency of the reinforced filter layer was slightly lower when the droplets were larger than 0.8 microns. Under the identical inlet concentration conditions, the treated filter material had greater number of droplets with a diameter of more than 0.8 microns than the untreated filter layer. It is believed that the liquid viscosity, surface tension and other physical properties were changed as a result of the mixing of the adhesive additives with the captured liquid droplets. This complicated the drainage process, and resulted in more re-entrainment of droplets. This conclusion can be verified by observations during the experimental process. The color of the liquid obtained from the reinforced filter layers was more turbid, and bubbles were generated at the outlet side of the layers. In addition, this phenomenon would be improved through placing a hydrophobic drainage layer after the reinforced fibrous layer. In general, application of adhesives to the filter appeared to enhance the structural strength of the filter material, and caused no obvious negative effects to the filtration efficiency. This method could be used for the industrial manufacture of filter cartridges after optimizing the processing technology. 3.3 Hydrophobic drainage layer A gas-liquid filtration process includes a complex process for liquid droplet interception, coalescence, migration and entrainment. The liquid outside the filter cartridge would form a large number of bubbles from the liquid film under the action of the gas drag force, and the bubbles gradually grow and burst into lot of droplets at the outlet side of the filter cartridge, then the captured liquid was re-entrained with gas flow.
12
Related research of liquid aerosol filtration by hydrophobic filter materials, has shown that hydrophobic filter material can significantly reduce liquid saturation and the risk of liquid re-entrainment in the gas stream [22-24]. Therefore, to further improve the properties of the subject filters, a layer of hydrophobic polyester fiber was placed in the filter after the coalescence layer and transition layer. A comparative test of the gas-liquid filtration performance was then conducted under the same experimental conditions previously employed, to determine the effect of the addition of this hydrophobic layer. The results of the tests are shown in Fig. 10. As shown, it appeared that the new-type of coalescing filter cartridge reached the steady state earlier, and the pressure drop of C2 was only 25% of C1 under the same operating conditions. The 0.3 micron droplet filtration efficiency reached 99.99%, indicating that the filtration performance of C2 was significantly improved. 3.4 Performance matching of the two-stage filter cartridges The degradation of filtration performance of the pre-filter cartridge was reduced by adjusting the structure and parameters of filter medium. More specifically, the design of a new type pre-filter cartridge was based on three aspects as shown in Fig. 11. The pleated layer with a high dust holding capacity was designed as the first layer at the inlet side to cope with the larger contaminant particles, with special attention to the operating conditions of pigging and batch processing. A group of fine filter layers was placed in the cartridge to capture the fine particles that penetrated the first layer, to ensure filtration precision. A frame was set at the inlet side to ensure that the filter would not be crushed when the pressure drop was large and the cartridge frame material was 316L stainless steel to avoid the corrosion in acid environments. Eventually, the pressure drop across A2 decreased by 50% over that of A1. This was because by the increase in droplets with a diameter of less than 2 micron that
13
penetrated through A2 and were captured by the coalescing filter. Therefore, the operating load of the two-stage filters received a reasonable allocation of droplets. Fig. 11 shows the pressure drop for the combination of A2 and C2. The pressure drop in both A2 and C2 gradually increased and the total pressure drop decreased. The droplets with a diameter of more than 2 micron were nearly all removed from the gas stream after A2 and the outlet concentration of droplets from C2 was relatively low. However, the practical effects of the improved filter cartridges remained to be demonstrated in actual field application. 4. Application of improved filter cartridges A combination unit in a natural gas purification plant in southwest China was chosen as the site to conduct the field tests on the improved filter cartridges. This plant had two series units as a combination unit. The conventional filter cartridges were separately installed as the control group and the improved filter cartridges were installed as the trial group. Process data were recorded during the test, including the inlet gas pressure, the gas flow from the filtration unit, the pressure drop and liquid drainage volume of the pre-filter and coalescing filter. The data were accumulated for 800 hours of testing and were selected for a comparative analysis of the filters. The inlet pressure was stable at about 8 MPa during the field test and the gas flows of the two series were nearly the identical. For the trial group, the pressure drop of the two-stage filters have risen to 15 kPa and 4 kPa, while the values for the control group showed no obvious change. The increased pressure drop was due primarily to the accumulation of solid particles that clogged the filter layers. Therefore, the increase in the pressure drop between the two groups showed that the trial group filter exhibited better filtration performance. During the 800 hours of testing, the liquid that accumulated in the control group filters 14
was drained 15 times for a cumulative drainage volume of 0.95 m3. By contrast, the trial group filters were drained 28 times with a drainage volume of 2.03 m3. These results indicated that the trial group exhibited more than twice the liquid intercept capability of the control group. These results showed that the improved filter combination had significantly improved filtration performance for both solid particles and liquid droplets. Also, the two-stage filter exhibited a synchronous increase in filtration, which achieved the goal of performance optimization and matching of the two-stage filtration unit in a natural gas purification plant. 5. Conclusions The feed gas in a natural gas purification plant contains high concentrations of hydrogen sulfide, high liquid content and experiences fluctuating working conditions. Factors including operating conditions, filter performance and operation costs, must be comprehensively considered when choosing the type and materials of the filter cartridges used to clean up the gas stream. This reported work focused on the problems existing in natural gas filtration plants where the pressure drop of a coalescing filter and a pre-filter are mismatched. Comprehensive analysis of the problem showed that in the two-stage filtration unit that was designed without matching the performance of the pre-filter and coalescing filter, the coalescing filter was individually incapable of removing sub-micron particles. Methods were presented for improving the filtration efficiency and stability of coalescing filter cartridge. To obtain low resistance and highly efficient operation in a two-stage filtration unit, reasonable cooperation and performance matching were examined, together with optimization of the performance of the pre-filter cartridge and the coalescing filter cartridge. Based on the experimental results of these studies, especially the method of improving the coalescing filter cartridge and performance matching of two-stage filter 15
cartridges, this present study provided a reference for designing high-pressure and corrosive gas filtration system. Acknowledgement This work was supported by China National Natural Science Foundation (No. 51376196) and China National Science and Technology Major Project (No. 2016ZX05017). References [1]
Y. Alcheikhhamdon, M. Hoorfar, Natural gas quality enhancement: A review of the conventional treatment processes, and the industrial challenges facing emerging technologies, J. Nat. Gas Sci. Eng., 34 (2016) 689-701.
[2]
D. Engel, M.H. Sheilan, Seven Deadly Sins of Filtration and Separation Systems in Gas Processing Operations, in: Y. Wu, J.J. Carroll, Q. Li (Eds.) Gas Injection for Disposal and Enhanced Recovery, John Wiley & Sons, Inc., Beverly, 2014, pp. 227-241.
[3]
P. Cézac, J.P. Serin, J.M. Reneaume, J. Mercadier, G. Mouton, Elemental sulphur deposition in natural gas transmission and distribution networks, J. Supercrit. Fluid., 44 (2008) 115-122.
[4]
E. Alhseinat, P. Pal, A. Ganesan, F. Banat, Effect of MDEA degradation products on foaming behavior and physical properties of aqueous MDEA solutions, Int. J. Greenh. Gas Con., 37 (2015) 280-286.
[5]
P. Pal, A. AbuKashabeh, S. Al-Asheh, F. Banat, Role of aqueous methyldiethanolamine (MDEA) as solvent in natural gas sweetening unit and process contaminants with probable reaction pathway, J. Nat. Gas Sci. Eng., 24 (2015) 124-131.
[6]
Y. Ke, B. Shen, H. Sun, J. Liu, X. Xu, Study on foaming of formulated solvent UDS and improving foaming control in acid natural gas sweetening process, J. Nat. Gas Sci. Eng., 28 (2016) 271-279.
[7]
C.R. Pauley, R. Hashemi, S. Caothien, C.R. Pauley, S. Caothien, Ways to control amine unit foaming offered, Oil Gas J., 87 (1989) 67-75.
[8]
T. Hu, G. Xiong, J. He, J. Yin, X. Peng, Prevention and control measures for 16
foaming in amine desulfurization unit, Natural Gas Industry, 29 (2009) 101-103. [9]
Z. Wang, C. Yang, L. Zhu, Corrosion analysis and inhibition studies in the process of natural gas wet desulfurization, Eng. Fail. Anal. , 44 (2014) 66-73.
[10]
F. Li, G. Sun, Q. Zhang, J. Long, Corrosion behavior and prevention of gas sweetening unit in natural gas processing plants, Natural Gas Industry, (2009) 104-106.
[11]
B. Li, Z. Ji, X. Yang, Evaluation of gas-liquid separation performance of natural gas filters, Pet. Sci., 6 (2009) 438-444.
[12]
M.D.M. Innocentini, E.H. Tanabe, M.L. Aguiar, J.R. Coury, Filtration of gases at high pressures: Permeation behavior of fiber-based media used for natural gas cleaning, Chem. Eng. Sci., 74 (2012) 38-48.
[13]
M. Azadi, A. Mohebbi, S. Soltaninejad, F. Scala, A case study on suspended particles in a natural gas urban transmission and distribution network, Fuel Process Technol., 93 (2012) 65-72.
[14]
M. Azadi, A. Mohebbi, F. Scala, S. Soltaninejad, Experimental study of filtration system performance of natural gas in urban transmission and distribution network: A case study on the city of Kerman, Iran, Fuel, 90 (2011) 1166-1171.
[15]
C.R. Pauley, D.G. Langston, F.C. Betts, Redesigned filters solve foaming, amine-loss problems at Louisiana gas plant, Oil Gas J., 89 (1991) 50-55.
[16]
Z. Liu, Z. Ji, X. Wu, H. Ma, F. Zhao, Y. Hao, Experimental investigation on liquid distribution of filter cartridge during gas-liquid filtration, Sep. Purif. Technol., 170 (2016) 146-154.
[17]
C. Chang, Z. Ji, C. Liu, F. Zhao, Permeability of filter cartridges used for natural gas filtration at high pressure, J. Nat. Gas Sci. Eng., 34 (2016) 419-427.
[18]
M. Finšgar, J. Jackson, Application of corrosion inhibitors for steels in acidic media for the oil and gas industry: A review, Corros. Sci., 86 (2014) 17-41.
[19]
A.M. Saleh, H.V. Tafreshi, B. Pourdeyhimi, Service life of circular pleated filters vs. that of their flat counterpart, Sep. Purif. Technol., 156, Part 2 (2015) 881-888.
[20]
A. Charvet, Y. Gonthier, E. Gonze, A. Bernis, Experimental and modelled efficiencies during the filtration of a liquid aerosol with a fibrous medium, Chem. Eng. Sci., 65 (2010) 1875-1886. 17
[21]
P. Contal, J. Simao, D. Thomas, T. Frising, S. Callé, J.C. Appert-Collin, D. Bémer, Clogging of fibre filters by submicron droplets. Phenomena and influence of operating conditions, J. Aerosol Sci., 35 (2004) 263-278.
[22]
B.J. Mullins, R. Mead Hunter, R.N. Pitta, G. Kasper, W. Heikamp, Comparative performance of philic and phobic oil‐ mist filters, AIChE J., (2014).
[23]
C. Chang, Z. Ji, F. Zeng, The effect of a drainage layer on filtration performance of coalescing filters, Sep. Purif. Technol., 170 (2016) 370-376.
[24]
R. Mead-Hunter, A.J.C. King, B.J. Mullins, Aerosol-mist coalescing filters - A review, Sep. Purif. Technol., 133 (2014) 484-506.
18
Figure Legends: Fig. 1 Feed gas filtration unit in natural gas purification system. Fig. 2 Filtration performance evaluation system for two-stage filter cartridges. Fig. 3 Main reason for failure of the fibrous medium. (the schematic diagram was excerpted from the literature published by Pall Corporation, USA) Fig. 4 Improved coalescing filter cartridge design. Fig. 5 Influence of the protecting layer on the pressure drop of coalescing filter cartridge C1. Fig. 6 SEM micrographs of fiber net with/without adhesive. Fig. 7 Influence of adhesive on initial pressure drop of filter layers. Fig. 8 Influence of adhesive on pressure drop of filter layers during gas-liquid filtration. Fig. 9 Influence of adhesive on filtration efficiency of filter layers. Fig. 10 Comparison of gas-liquid filtration performance of coalescing filter cartridges. Fig. 11 Structure design of pre-filter cartridge. Fig. 12 Pressure drop evolution for the combination of A2 and C2.
19
ΔP
Clean gas Feed gas
Pre-filter A
ΔP
Coalescer C ΔP
Standby Pre-filter B
Fig. 13 Feed gas filtration unit in natural gas purification system.
20
Compressed air
Upstream sampling
Downstream sampling
Mixing chamber
Liquid
Coalescing filter
Pre-filter
Protect filter PID control
Spraying nozzle Aerosol particle size spectrometer
PID control
Flowmeter Valve V-3
T&H
Air Conditioner
Liquid Gatherer
Air
Blower
ΔP
ΔP
Midstream sampling
Inlet filter
Liquid Gatherer
Fig. 14 Filtration performance evaluation system for two-stage filter cartridges.
21
a. Contaminant unloading
b. Structure damage
104μm
573μm
Fig. 15 Main reason for failure of the fibrous medium. (the schematic diagram was excerpted from the literature published by Pall Corporation, USA)
22
Coalescence layer Transition layer Adhesive particles Outer frame Protective layer
Inner frame
Drainage layer
Fig. 16 Improved coalescing filter cartridge design.
23
Original Coalecing filter C1 C1 with pleated protecting layer
Pressure drop (kPa)
2.0
1.5
1.0
0.5
0.0
0
20 40 2 Dust loading per unit filter area (g/m )
60
Fig. 17 Influence of the protecting layer on the pressure drop of coalescing filter cartridge C1.
24
a. Clean state
b. With adhesive
Fig. 18 SEM micrographs of fiber net with/without adhesive.
25
1.0
Pressure drop (kPa)
0.8
0.6
0.4
Untreated filter layers Reinforced filter layers
0.2
0.0 0.04
0.06
0.08
0.10
0.12
0.14
Face velocity (m/s)
Fig. 19 Influence of adhesive on initial pressure drop of filter layers.
26
4.0
Pressure drop（ kPa）
3.5 3.0 2.5 2.0 1.5 1.0
Untreated filter layers Reinforced filter layers
0.5 0.0
0
1
2
3
4
5
6
Time (h)
Fig. 20 Influence of adhesive on pressure drop of filter layers during gas-liquid filtration.
27
Filtration efficiency (%)
100.0 99.9 99.8 99.7 99.6
Untreated filter layers Reinforced filter layers
99.5 99.4
0.3
0.5
1 2 Droplet size (μm)
4
8
Fig. 21 Influence of adhesive on filtration efficiency of filter layers.
28
16
8 kPa
Pressure drop (kPa)
8 4
2 kPa
2 1 0.5
Coalescing filter cartridge C1 Coalescing filter cartridge C2
0.25 0
1
2
3
4
5
Time (h)
a. Pressure drop evolution
Cumulative efficiency (%)
99.9999 99.999
99.99%
99.99
99.9%
99.9 99.5 99 98
Coalescing filter cartridge C2 Coalescing filter cartridge C1
95 0.3
0.5
1
2
4
8
Droplet size (¦ Ìm)
b. Filtration efficiency at steady state Fig. 22 Comparison of gas-liquid filtration performance of coalescing filter cartridges.
29
a. Original type Single medium without frame
b. New type High dust holding layer Fine filter layer
Stainless steel inner frame
Fig. 23 Structure design of pre-filter cartridge.
30
3.0
Pre-filter cartridge A2 Coalescing filter cartridge C2
Pressure drop (kPa)
2.5 2.0 1.5 1.0 0.5 0.0
0
2
4
6
8
10
12
14
Time (h)
Fig. 24 Pressure drop evolution for the combination of A2 and C2.
31
Highlights (1) The failure of the coalescing filter mostly results from the deformation of fibrous structure and the fracture of the fibers. (2) Application of adhesives to the filter could enhance the structural strength of the filter material. (3) The two-stage filtration unit was designed with consideration of the performance matching of the pre-filter and coalescing filter.
32