Originally appeared in: February 2007, pgs 109-116. Used with perission. An online sulfur onitoring syste can iprove process balance sheets A Canadian gas processor used this technology to eet environental regulations S. Wang, B. Harris and S. Mackenzie, Gas Liquids Engineering Ltd., Calgary, Alberta, Canada Sulfur eissions fro sour gas processing facilities are a serious concern to the environent and personal health. Governents have ipleented strict guidelines for onitoring and reporting eissions created by such processing facilities. In Alberta, Canada, the Energy Utilities Board (EUB) requires gas producers to provide a report that clearly indicates how uch sulfur is brought into a plant and how uch is either converted to eleental sulfur or injected as acid gas. The report is known as the S-30 Sulfur Balance Report. Producers are required to produce a balance fro the inlet to outlet on sulfur tonnage with a 10% uncertainty. This is equal to ±5% of the total balance for a onthly average. Many gas processing plants suffer iprecision in evaluating sulfur ibalance over a period of tie due to the difficulties in accurately easuring and calculating the sulfur balance. A real-tie online sulfur onitoring syste (RTOSMS) was developed for acid gas processing plants to assist operations and anageent by providing real-tie onitoring and reporting of sulfur ibalances. Sulfur onitoring before RTOSMS. A grab saple fro necessary strea was taken once a onth and its coposition would then be used until the next sapling period. Daily sulfur ibalances for that onth were then evaluated using these gas copositions, as well as the strea average daily standard voluetric flowrates. This ethod, however, produced unsatisfactory results and created errors in anaging sulfur balance. The ain reason this practice is insufficient is due to one or ore factors: Inlet and outlet strea flow changes and its easureents Sulfur, % 18 14 12 10 8 6 4-2 02-4 -6-8 -10 Fig. 1 GCs Fig. 2 Data 6:00 a Sulfur ibalance history trend, %. Coposition Coputer controls Changes in each strea s gas coposition and its easureents Changes in strea flow conditions and its easureents Sour gas plant processing cycle tie. Fig. 1 shows the historical sulfur ibalance trend for a 24-hr period (Pouce Coupe gas plant, February 4, 2005). Although the daily average ibalance is 2.5386%, in 12:00 a 6:00 a 12:00 a Syste RTOSMS structure and data flow ap. Sulfur onitoring Data anageent Flow calculation Sulfure balance Flare alar Report and record History review Results Real-tie Monitoring Recording S-30 report Flare report Manageent Docuentation stantaneous easureents varied between 10% and 17%. Thus, taking a grab saple anytie that day would produce different results. Also, taking grab saples at periodic ties have shown errors of ±20% or ore fro real-tie sapling. Errors associated with grab saples are: 1. The tie between sapling and analysis. Most sour gas production plants are
Water content, ol fraction 0.50 0.45 0.40 0.35 0.30 0.25 0.20 0.15 0.10 0.05 0.00 Fig. 3 Sulphur ibalance, % 9 7 5 3 1-1 -3 H 2 S = 0.458 ol fraction CO 2 = 0.458 ol fraction H 2 O = 0.084 ol fraction M w = 37.3 kg/kol Z = 0.9921 0 10 20 30 40 50 60 70 80 90 100 Teperature, C located in rural areas and analysis labs are located in ajor cities. In soe cases, the saples can be stored in a vehicle for ore than three days before analysis. 2. Saples are taken in areas where, in case of leakage, the technician can safely evacuate away fro the saple point. Because of this, saple lines are usually installed with long runs of tubing. This tubing is usually not heat traced and produces a cooling effect on process saples (30 C). This results in liquid condensation and saple distortion. 3. Process copositions can vary fro hour-to-hour and day-to-day. In the event H 2 S = 0.34 ol fraction CO 2 = 0.34 ol fraction H 2 O = 0.32 ol fraction M w = 32.3 kg/kol Z = 0.9924 H 2 S = 0.404 ol fraction CO 2 = 0.404 ol fraction H 2 O = 0.192 ol fraction M w = 35.0 kg/kol Z = 0.9887 H 2 O content (dry basis) of 50% H 2 S and 50% CO 2. 50 kpag 70 kpag 90 kpag 110 kpag 130 kpag 150 kpag -5 4:48:00 6:00:00 7:12:00 8:24:00 9:36:00 Tie Fig. 4 Sulfur ibalance correction by odified gases. Ibalance by raw copositions Ibalance by odified copositions Ibalance differential of a plant upset, it can take as long as an hour before the plant is balanced again. Taking grab saples at the inlets and outlets when there is a production upset can result in large differences in ass balance. To eliinate errors that are associated with grab saples, a real-tie approach is required utilizing technology that helps iprove accuracy and protect the workers and environent fro ishaps. RTOSMS structure. Online onitoring of sulfur ibalance and sour gas eissions is realized by continuously watching and analyzing flow, its related easureents and gas copositions for all involved strea. The plant operating syste provides the raw data used by the RTOSMS for process strea easureents. Gas chroatographs (GCs) are installed and can either be connected to the plant coputer or directly linked to the RTOSMS engineering coputer (Fig. 2). For the Pouce Coupe project, wireless counication was used to transit gas copositions fro each installed GC. This technology was chosen because it is relatively easy to install, ade the syste transportable, and provided cost effectiveness since there was an installed coputer control syste. The RTOSMS engineering station was then connected to the plant loop to obtain raw data fro the process easuring devices. GCs can be set up to analyze one strea or ultiple strea. The RTOSMS utilized one GC for the inlets and one for the outlets. This provided a workable syste and prevented the high-concentration hydrogen sulfide (H 2 S) strea fro containating the lower-concentration H 2 S strea. Process data analysis sequence. Data collected fro the GCs and plant coputer are organized on a database platfor. For the Pouce Coupe gas plant, a 10-sec. scan rate was chosen. A saller scan rate can be used but results in needing ore database storage. The 10-sec. scan rate was also sufficient in obtaining correct flow easureents. For a valid scan, the syste follows this sequence to verify and calculate results: 1. Gas coposition odification 2. Calculation of the standard voluetric orifice flowrate 3. Sulfur ibalance calculation and alar onitoring 4. Flare event onitoring 5. Data reporting, recording and anageent In the event that the process causes an incorrect result fro either a process flow condition or disturbance, the RTOSMS autoatically reports this condition and ignores the value. This is very iportant for analyzing strea that are very heavily saturated with (H 2 O), because ost acid gas strea operate near or in a two-phase region. Not being able to calculate when the acid gas strea is near or on the verge of becoing a two-phase strea will result in erroneous readings. Gas coposition odification. Both the GC and grab saples only give gas coposition results on a dry basis, which eans that the H 2 O in the gas strea is not included. Neglecting strea
Table 1. Gas coposition odification Strea Inlet outlet 1 outlet 2 Teperature, C 18.8 29.8 25.4 Pressure, kpag 3,977.6 57.2 56.9 Gas Mol % Mol % Mol % coponent Raw Modified Raw Modified Raw Modified H 2 O 0.0647 2.8489 2.2133 H 2 S 0.8462 0.8456 48.2347 46.8606 43.5511 42.5872 CO 2 1.8299 1.8287 51.2617 49.8013 55.9694 54.7307 Methane 90.4801 90.4216 0.4664 0.4531 0.4314 0.4219 Ethane 4.5467 4.5438 Propane 1.4092 1.4084 i-butane 0.2189 0.2188 0.0052 0.005 0.0043 0.0042 n-butane 0.3767 0.3764 0.0142 0.0139 i-pentane 0.0863 0.0862 0.0292 0.0284 n-pentane 0.1033 0.1032 0.0028 0.0027 0.0295 0.0289 Hexane+ 0.1026 0.1026 The strea calculated flowrate and sulfur production Flowrate, 1167.483 1162.8100 10.2261 10.2194 11.8376 11.8125 e 3 3 /d Sulfur, 13.3401 13.3324 6.6881 6.4933 6.9903 6.8211 ton per day (tpd) Sulfur ibalance 2.4738 0.1348 Online onitoring page. H 2 O content results in calculation discrepancies of the flowrates, sulfur content and plant sulfur ibalance (Table 1). Fig. 3 shows how the saturated H 2 O content of a 50% H 2 S and 50% carbon dioxide (CO 2 ) ixture is changed due to these conditions. A quick calculation of the sulfur content (dry basis) at 50 C would result in the difference shown in Table 2. A ethod developed especially for H 2 O and acid gas equilibriu was introduced to adjust the Fig. 5 strea s gas coposition according to its flowing condition (pressure and teperature) and original ol% of gas coponents fro the GCs. 1,2,3 A coplex algorith is used to calculate the H 2 O content and produce a strea ixture that is a ore accurate representation of the process strea being used. Table 1 illustrates the results of flowrates and sulfur in the strea with and without odifications. The odification perforances on different strea vary. Generally, the calculated sulfur ibalance will decrease by gas coposition odification because there is actually less sulfur being produced than what is expected. The ethod s contribution to correcting sulfur ibalance varies fro tie-to-tie due to the flowing conditions and gas copositions (Fig. 4). It also adds about a 5% decrease in the sulfur ibalance for this case. Fro both theories and practice, a better and ore reasonable calculation of flow easureents and evaluating sulfur ibalance can be derived by gas odification and the RTOSMS. This is done by taking the H 2 O content into account for gas copositions. Sulfur ibalance calculations. The sulfur ibalance for a sour gas plant is calculated as: out in S = () % 100 1 s out For each gas strea, the ass flowrate of sulfur is coputed fro the strea gas flowrate and the ol fraction of the H 2 S in the strea. 4 s = 1.35592Q y HS 2 ( 2) H 2 O content. The H 2 O ass flowrate can be calculated by: gas = MW MW gas y ( 3) Flow eters and flow easureents. Careful selection of easuring devices and installation ust be considered when easuring strea with high acid gas concentrations. The reason for this is not only H 2 O content but also flow conditions. Due to the low pressure and viscosity of the acid gas strea, it is iportant that correct flow easureents are taken. It requires that the easuring devices produce very low-pressure restrictions and also prevents H 2 O fro obstructing or interfering with the flow easureent device. Both vortex and senior orifice eter runs were used for the plant. However, before they could be used, installation corrections were required. Coon errors found in easureents were noticed and changes ade. For a vortex eter, the standard voluetric flowrate is: ρpt, Q = Q PT, ρ ( 4) For the orifice easureent, the API Natural Gas Fluids Measureent Standard 5 is followed exactly to calculate the ass flowrate:
Table 2. H 2 O content calculation for a single strea Strea coposition, ol % Flowrate, e 3 3 /day Sulfur produced, tons % error 50% H 2 S with no H 2 O content 10 6.78 8.4 at 50 C and 50 kpag (dry basis) 45.8% H 2 S with H 2 O content 10 6.21 at 50 C and 50 kpag Fig. 6 History review page. 2 q = N C E Y d ρ, P ( 5) 1 d v P T Eq. 4 is used to convert the ass flowrate to the standard voluetric flowrate required by Eq. 2 for sulfur ass flowrate. Gas density estiation has a great ipact on equation accuracy. The Aerican Gas Association detail characteristic ethod 6 is used here and can generally be expressed as: ρ PT = f P, T, x, ( i ) ( 6) Real-tie onitoring. Continuous online onitoring for the plant study included these strea: The inlet strea flow cobines with one inlet strea off the inlet separator and a strea generated fro the stabilizer overhead. The outlet strea flow consists of parallel acid gas strea that are generated fro two processing plants. Each acid gas strea is equipped with one acid gas flare strea. To easure strea coposition, one GC is used to easure the inlets and one GC is used to easure the outlet strea. Analysis tie for each strea varies fro 3 in. to 5 in. depending on GC configuration. The RTOSMS scans the online run-tie data fro the database every 10 seconds. The data status check is perfored first to see if there is any error in the data acquisition syste. If successful, the data status checks this scan s data which will be saved to database history for later detailed analysis. This works exactly the sae as the online onitoring; otherwise, this scan will be discarded and the progra will wait for the next update. Fo r e a c h g a s strea, the calculation is applied to the GC gas coposition at the easured flow teperature and pressure aking the H 2 O content adjustent to gas coposition. The proper flow calculation procedure will be initiated according to the floweter and its installations. Finally, the strea s sulfur and H 2 O ass flowrate will be calculated using Eqs. 2 and 3. After finishing the calculations for all strea, the sulfur ibalance can be finalized with Eq. 1. After each scan, the data and results will be sent to the online onitor report page (Fig. 5). The sulfur ibalance will be shown in alar color if its absolute value is larger than 5%. A trend chart, capable of various tie periods, is provided to show historical results. The error-essage window on the coputer screen shows any error essage found during processing. The ost coon error essage is the twophase flow error. In addition to instant results, the hourly, daily and onthly analysis suaries will be generated and saved and autoatically added to the database. The results can be downloaded to other coputers and reviewed, or reports can be generated fro this coputer. Historical review. The hourly, daily and onthly suaries saved in the database are anaged by the tie tree in the history review page (Fig. 6). This facilitates easy access to existing reports or suaries. The RTOSMS autoatically checks to ensure data has been collected for a 24-hr period. This is indicated by a red check ark beside each day. In the event that data is lost or not reported, the progra treats this as lost eission onitor tie which is reported onthly. This tool was added for properly anaging eission control devices and it prevents any eissions devices fro incorrectly recording false inputs. An existing suary and report ite can be overwritten or a new ite can be added for a specified day or onth anually. The selected daily or onthly report (S-30) can be exported to a spreadsheet and then printed or E-ailed depending on requireents. Fig. 7 shows the autoatically filled S-30 for for the Pouce Coupe gas plant, which was generated at the beginning of the onth fro the onitoring software. Flare event onitoring. It has becoe ore iportant, in the past few years, that producers anage and record flaring incidents. The RTOSMS can accurately easure and record acid gas flare events. Once a flare event happens, the startand end-tie, axiu and average standard voluetric flare flowrate and sulfur flared will be calculated and recorded to the database. All flares are listed on the flare report page with the latest at the top, as shown in Fig. 7. The operator can navigate in the record table and attach notes to the flares. This inforation can then be viewed by the operators and used to fill out the required flaring report. The flare report can be generated at any tie for the selected events and is capable of daily, onthly or yearly reporting. The reports will total the duration and volues to help anage flare events (Fig. 8). What has been learned? The RTOSMS has been working at two Duke Energy Midstrea Services gas plants for ore than one year. The syste s perforance deonstrates that this is the right solution to help assist in the daily operation of a sour gas processing plant. Syste benefits include: 1. RTOSMS provides a reliable and accurate way to evaluate the sulfur ibalance for a sour gas processing plant by continuously analyzing sulfur ibalance with the latest technologies. 2. This syste records all the detailed operation inforation about the plant electronically. It also autoatically generates daily and onthly (S-30) reports for operations and anageent.
Fig. 7 3. RTOSMS also provides an accurate ethod for easuring and recording acid gas flare events. This will help operators find out why flaring events occurred and hopefully help reduce such events by correcting the proble. 4. RTOSMS can be set up to onitor ultiple plants. Data can be stored on the plant site and retrieved a distance away. For this study, the sour gas plants were located in Canada s Grand Prairie region and the data was sent to Calgary where the reports were generated. HP Sybols: NOMENCLATURE S % Gas plant sulfur ibalance, % Gas plant inlet strea total sulfur in rate, ton/day (tpd) Gas plant outlet strea total sulfur out rate, tpd Flare report page. Sulfur strea ass flowrate, tpd Q Standard (@101.325 kpa, 15 C) voluetric flowrate of the gas, 1,000 standard cubic eters/day (S 3 /d) y H2 S H 2 S ol fraction in gas H 2 O ass flowrate, tpd gas Gas strea ass flowrate, tpd MW Mol weight of H 2 O, kg/kol MW gas Mol weight of gas, kg/kol y Mol fraction of H 2 O in gas Q PT, Actual voluetric flowrate of gas easured by vortex eter, 1,000 S 3 /d P Flow pressure, kpa T Flow teperature, C q Orifice ass flowrate easureent, kg/s N 1 Unit conversion factor, 1.11072 for SI units C d Discharge orifice plate coefficient E v Velocity of approach factor Y Expansion factor d Orifice plate bore diaeter calculated at flowing teperature, DP Orifice differential pressure, Pa x i Gas coposition odified by the RTOSMS. Greek letters: r P,T Gas density at flow condition, kg/ 3 r Gas density at standard condition (101.325kPa, 15 C), kg/ 3 Subscripts: s Sulfur in Inlet strea out Outlet strea H 2 S Hydrogen sulfide H 2 O H 2 O gas Gas P,T At flow pressure and teperature Mass flowrate ACKNOWLEDGMENTS The authors would like to thank John Carroll at Gas Liquids Engineering and Greg Rau at Duke Energy Midstrea Services for supporting this study. LITERATURE CITED 1 Carroll, J. J., Phase equilibria relevant to acid gas injection, Part 1 Non-aqueous phase behavior, Journal of Canadian Petroleu Technology, June 2002. 2 Carroll, J. J., Phase equilibria relevant to acid gas injection, Part 2 Aqueous phase behavior, Journal of Canadian Petroleu Technology, July 2002. 3 Carroll, J. J., The content of acid gas and sour gas fro 100 F to 220 F and pressures to 10,000 psia, 81st Annual GPA Convention, Dallas, Texas, March 11 13, 2002. 4 EUB Measureent Guide, Alberta Energy and Utilities Board, April 2004. 5 Copressibility Factors of Natural Gas and Other Related Hydrocarbon Gases, AGA Transission Measureent Coittee Report No. 8, Aerican Petroleu Institute MPMS Chapter 14.2, Gas Research Institute, Noveber 2003. 6 Manual of Petroleu Measureent Standards, Third Edition, Aerican Petroleu Institute, Septeber 1990. BIBLIOGRAPHY The Oil and Gas Conservation Act, Legislative Assebly of Alberta, RSA 2000. Shouxi Wang has over 15 years of cobined experience in engineering and software developent for oil and gas handling, and transportation. He holds a PhD in petroleu engineering, an MS degree in echanical engineering and a BS degree in oil and gas storage and transportation fro Southwest Petroleu Institute, China. Dr. Wang has extensive expertise and experience in the software developent of pipeline network siulation, pipeline leak detection and engineering applications. Bert Harris has 15 years of experience in the oil and gas industry with extensive experience in instruentation and analytical equipent. Along with that, Mr. Harris has worked on process design and plant optiization projects. He has a BS degree in cheical engineering fro the University of Calgary, Canada. Stuart MacKenzie has been a cheical engineer with Gas Liquids Engineering for five years. He has anaged a nuber of natural gas and oil production projects. Mr. MacKenzie is currently the lead engineer for sulfur onitoring research (MassTrak) and the software developent departent. Article copyright 2008 by Gulf Publishing Copany. All rights reserved. Printed in U.S.A. Not to be distributed in electronic or printed for, or posted on a Website, without express written perission of copyright holder.