Home

An Intelligent Portfolio Management Approach. Final Report

1. i Su sl pee il 211 Tue salir ll dl aa ume m penes ACTIVE Sires BTHELL Wellhead Wellhead lt Make gt install Date 00000 Fi dupli 10 Wellnead Components Dos Istavruxe Tyne Mukz Sizej WP inu Operetve Leaking Rasincred Comment 0 anma wawe Ba J EM I gt Gae Bardo 3000 Oe 80 0 0 Downhole C mpanents Dates 18580 are appre b ui pe aret o m ps em mr NK Tog eee rr 1913 557968 57311913 m 11968 3000 sens 11971505 vue el Ip TE CN 42884 81 1 1 iue 1 35 32 31 1 31 34 ING rat Ce 19 920 1 Qrorox23 LUCAS Report dale y February 22 2005 Page tat Perforations Due regm emm snot Type cuv ype 5218 note PO sum rerom se Date Top Tt Shot Tyne Gun Sz in shots String Per C
2. 5 Sec sy Treciuma bpi Type of Prssteres iii Oper Eli Feine minc Pied Une Facog Mod Acus P 1 wa Gallons p am E iun wt e 1 aT ri Feet A B r goir jae uox r um nain p Pu Ove Fur Gaps TET Oe hn be La alice ga i LITET Hes Fin iL bh Pind Inset pele ee ee Gaases Foie wapa Gasses LUELE a_i d a 7 nR raga inert Gas Cancepiratean was Bra page Stee Bete r pet ol Average Men Gan Concentrator Fate was SCF Cal m n 0 pes hartge at EE Ua Beers FRE Dit n ata Cee Oak WEN Agent SandiMesh D d dai ingr Propping Agent Le sh 1 d mh TE ETT E MEN A mourn F waant ai LL Concentration He Sand amp ha simum 5 Avg Ome Propping Town Flow afar Pract i ae Freee ELIO 03 Agents Aug ee JEL
3. NENNEN __ Dx ee ee d Anderson Bolander UC AS ES 340056 ainen RN Eo E pc m ii rmm 8 peus i40 Mess gram oe ee Fig1 Well bore data retrieved from a file 14 The tables contained many minor mistakes like wrong Well API number length and spelling mistakes A picture of this correction is shown below gar TS KROLIT ET AL 19912 SHAFFER al 2921 G HINER 2 03926 JOHN BUSLER 1 T3829 D KICK 1 3951 E GUTHRIE df T2922 J TRUMECWER 41 3835 J ANDREWS 32 3939 MAURER T3845 RUSSEL J UFER ET Ux ise C SMITH 1 13050 i PLIRVINE 1 T3253 J ROWE 1 T3953 amp MARY BREMNETUHL T5867 J MANG Hi 3950 J amp E PARR T3872 J H ROWE 42 3976 W amp APPLEGATE 1 amp C GUTHRIE df 39523 E amp OSWALD i 5995 5 E MCKENLEY aes amp APPLEGATE 22 T4008 amp M MCGUIRE sl 34 005 93897 34 005 600620000 34 005 93921 34 005 602650000 4 005 93929 34 005 93931 94139601 190000 34 0533835 34139601230000 4138201580000 TAUTA 34 05 9969 34 138 83372 34 139 93976 34 005 93978 34138601590000 34 05 9599 34 139 93997 4005216200000 SPECIAL ACTIVE 52 3058 17 245 52 371 924 82 2355 12 5223611 o Wrong API Number length ACTIVE ACT
4. Injection Rate bbls mn 3 Total Fluid bbls 23897 Summary Total bbls Flush bbls Sand Quantity Ibs Sand Type Well Test 2000 1 000 Break dawn Pr psi ISIP psi Date 8 2 956 MMOD eee Alcohol bbls 2 4 Cushion bbls 20 Previous Stimualtion Next stimulation Last Stimualtion Fig44 Stimulation tab To browse between different Stimulations Previous 5timualtion Mext Stimulation Last Stimualtion To move to the first stimulation previous stimulation next stimulation amp the last stimulation in the record click on the button assigned to it The stimulations are sorted in ascending order according to stimulation date for current well 62 WELL TEST wellbore Completion Stimulation Well Test well est 3 of 9 Time hrs Field Pr psi Pr psi Hate McD Well Test First Reading 05 EE BE cco 2029 Field Name LUICAS Second Heading as 1020 55 EIN Well T0803 Third Reading 075 ECH gu 1420 Test Date 12 28 1998 Extended Time 2 ams oa Test Multi Point kh md ft AAG NonDarcyCo etf DeltaP psi 250000 Rate 5 ADF Skin 3 n Value 0 630 PD Rate 1749 C Skin True Skin 2 5 C Value 0 65641 ADF 0 553 All Well Tests Well Test Analysis Previous WellTest Fig45 Well test tab To browse between different Well Tests First WellTest Pr
5. PD Rate ADF Skin C All well Tests Fig59 Well test Analysis Option in well test tab Peak day AOF Skin and all well test graph Select PD rate AOF Skin or All well Tests option button and then click on the Show button Show The following screens will appear according to the option selected Storage Technology Consortium STIMULATION amp WELLTEST DATA WELL 34 138 94075 Peak Day Rate Stimulation Peak Day Rate Date Min 5711 1917 Date Max 7 1 2005 Fig60 Show Chart Peak Day Rate 77 miis Consortium STIRMIULATION amp VWWELLTEST DATA WELL 34 139 94075 eee Stimulation E LL Ca E e L LL Date 122005 a V a Fig61 Show Chart Absolute Open Flow Gas stone jechinolopy amp WWELLTEST DATA WELL 3 34 138 34075 mm Skin Stimulation m E rra Date 8711 1917 Date 2 7 72005 Fig62 Show Chart Skin 78 2 Storage Technology Consortium STIMULATION amp WELLTEST DATA WELL 34 139 94075 em Open Flow Single FPoint Stimulation E E a pn Date Min 53171317 Date 71 2005 Fig63 Sh
6. Bog 2227 xD DATA dep of Boom QU G D s E Number of D ata Entries i de m ant QU Fig12 Data addition and refinement for Stimulation Data INITIAL DATA AVAILABLE ap ou oe se d y M WELL TEST DATA Well test data 1s the most extensive dataset that our R amp D team worked on It has the maximum amount of records nearly 3365 and 29 data types that control every aspect of a well test The data type selected for a well test representation consists of following 1 Well API Number 2 Field Name 3 Test Date 4 Test Type 5 Time 1 6 Field Pressure 1 7 Flowing Pressure 1 8 Rate 1 9 Time 2 10 Field Pressure 2 11 Flowing Pressure 2 12 Rate 2 13 Time 3 14 Field Pressure 3 15 Flowing Pressure 3 16 Rate 3 17 Time Extended 18 Field Pressure Extended 19 Flowing Pressure Extended 20 Rate Extended 21 kh 22 Skin 23 True Skin 24 Darcy Co efficient 25 Value 26 Value 2 Delta Pressure Squared 28 Peak Day Rate 29 Absolute Open Flow 30 Estimation of n C peak day rate amp absolute open flow Single Open flow Tests The values used for point 1 and 2 are from different well tests l Find AP log p p log p Pyy s BBP 2 OBP Where q is in MMcfD n log 4 log q 4 C Where 4 is in McfD p 5 AOF 1150 0 _ McfD 6 PDRate x 250 000 McfD Multi Point Test
7. Hange Erg un al iL 2 bb ky 024 2 LL 7 2 48 Sood 300 E DE 13 Afier 3 Tr 2 Ul m 2 Single Pa ini al Eric ind l blew Table 4 Calculation to determine the length of chromosome The length of chromosome came out to be 9 11 10 8 9 9 2 58 The GA characteristics that were used are shown in Table 5 These were the best but can be changed as desired to suit other neural nets in the future GA CHARACTERISTICS Crossover rate Mutation rate Population size of Generations 10 Next Generation criteria Top 30 ranked from previous generation Crossover criteria Top 25 has 75 chance of Crossover Table 5 CA characteristic There are two optimization methods made available in this software One is optimization just based on skin and other based on both skin and cost The optimization objective function 15 calculated using the following formula and minimizes this optimization objective function 47 eer ONERE Skin Shin Cost Cost Optimization Objective Function Mami zi x Cast weight in a x WELE HT m Dsl a Software compatibility and variability In the software user has been given many options to accommodate the particular situation that he has and data availability 1f different from the data that we have used to verif
8. o d STONGE OVS Hons RIE 8 5 0 Tesi After Acti pag MCF 2 Fig11 Different formats of Nitrogen Amount During the data entry different sign conventions and unit conversions were carried out as follows MEASURED JN UNITS USED IN QUANTITY RECORD DATABASE SAND 100 lbs 1 bbl 42 gal 2000 2005 ft 10 shots 2 shots ft 10 5 24 The following notations were used in place of different notations being used in the tables NOTATIONS WATER PETRO FOAM MITRO WATER T2 USED IN GEL FRAC SHOT FRAC DATABASE elled Petr gel Foam ohot Water 7 water Water NOTATIONS frac THAT Petro Gel Water WERE Fracture REPLACED Petro gel Water nitrogen fracture Water assist Water Fracture w N assist records of Nitro shots were discarded for this database as they have no stimulation parameters on record and are part of history now plus they also damage the well Above all they will tend to degrade the Neural Network 28 Analysis raw data vs refined data er er e 2 er un omm f me 3 DR mem s wem s 8 O 181 zr 202 365 255 32 Bus yo Bers EOL 32 32 32 H IMITILAL DATA AVAILABLE DATA AVAILABLE STIMULATION DATA 29 t 58 49 AS a d M ie FS
9. ASHLAND ASHLAND WERMILUON ASHLAND GREEW GREEW GREEN MONROE GREEW ASHLAND ASHLAND ASHLAND RICHLAND ASHLAND VERNILLIOM ASHLAND GREEN GREEN MONROE MONROE HANGER GREEN GREEN MOMROE MONROE GREEN MONROE ASHLAND ASHLAND RICHLAND RICHLAND ASHLAND ASHLAND ASHLAND RICHLAND RICHLAND AcHLANLU RICHLAND VERIMILLION ASHLAND MONROE GREEN OH OH OH OH OH OH OH OH OH OH OH OH OH OH OH OH OH OH OH RICHLAND ASHLAND mee Columbia Gas Transmissir Columbia Gas Tranamissit Columbia as Transmissic Columbia Gas Transmissit Columbia Gas Transmissic Columbia Gas Transmissit Columbia Gas Transmissir Columbia Gas Transmissit Columbia Gas Transmissic Columbia Gas Transmissit Columbia Gas Transmissit Columbia Gas Transmissic Columbia Gas Transmiseic Columbia Gas Transmissit Columbia Gas Transmiseic Columbia Gas Transmissit Columbia Gas Tranamissit Columbia Gas Transmissit Columbia Gas Tranamissit Columbia Gas Transrmissit Columbia Gas Transmissic Columbia Gas Transmissic 15 Analysis of raw data vs refined data 420 DATA FIELDS INTTHAL DATA AVAILABLE FINAL DATA AVAILABLE L1INITIAL DATA AVAILABLE WELLBORE DATA Field Name Lease Name Classification perator E z a eT 1430 430 427 429 430 430 430 430 430 430 WELLBORE DATA Number of D a Entries Fig3 Data addition and refinement for Well bore Data Total
10. 10543 ZADA GUTH ACTIVE 20472 31611 34 005 10552 10552 Ww C SCHA ACTIVE 71222 22444 5 34 005 10575 10575 ZELLA BITT ACTIVE 70972 2929 34 005 10557 10657 JOHN THOF ACTIVE 54472 30611 B 34 005 10676 10676 BERNARD ACTIVE 1 29933 23806 3 34 005 10795 10735 ROGER F ACTIVE 4883 A472 10 34 005 10898 10838 F REST E ACTIVE fared 20945 34 005 11029 11029 OHIO FUEL ACTIVE 2778 32777 34 005 20202 7359 HUNGEF ACTIVE 57694 33667 34 005 20370 8549 D KEYS ACTIVE 55917 3375 14 34 005 20373 8538 ESTHER ACTIVE 57861 33639 34 005 20388 8817 C C LEIBO ACTIVE 71806 31194 2 17 1 9947 HINES ACTIVE 72597 Type of Analysis Optimization Method Al wells GA based on Skin 3 C Single Well GA based on Cost And Skin Fig82 Candidate Selection main screen 95 When this module is loaded each row in this section of the grid represents a valid stimulation as shown in figure above Following is the criteria for valid stimulation selection Valid Stimulation It should have skin value before amp after stimulation Valid Perforation Perforation just before the stimulation Valid Completion The smallest size completion run before stimulation Valid Well test Well test having skin value just before or after the well test Two types of analysis can be done on the wells One option is to apply Genetic optimization on wells one at a time and the other 15 to
11. Candidate Selection from main screen Candidate Selection main screen adid Selecting a well for candidate selection Options to control Candidate Selection process Cost analysis Mod le babita Inputs that used to train the Neural Network Select the controllable parameters in optimization process Setup doxes If one of the Neural Net input is well test before stimulation it could setup here Optimization process for one Well c ccc eeeeeeee eee eeeeeeeeeeeees Optimization result for selected wells useeeeseeseeeeeeese Rank the optimization result based on delta skin in order to find the best candidate LIST OF TABLES Table 1 Draw down Test Results Table 2 Build up test results Table 3 Average Results Table 4 Calculation to determine the length of chromosome Table 5 CA characteristic 37 39 39 47 INTRODUCTION Each year Gas Storage operators spend hundreds of thousands of dollars to combat the inevitable decline in the deliverability of their production wells The decline in deliverability with time has two major contributors The first contributor 15 geology and reservoir characteristics that are uncontrollable parameters The second sets of parameters that contribute to the decline are associated with well damage that is addressed by well remedial operations such as work
12. Cost Import Net Nek Input Values 64 Characteristic Export the Grid To Excel Select Well TestType Fig84 Options to control Candidate Selection process Material cost can be changed by the user as the prices fluctuate These prices can be saved in the database by selecting Save button and Default values can be retrieved by selecting Default button The screen shot of material cost is shown below and price is just an estimate and can be changed by user 97 EU Material Cost Cost oe Water Injection bbls Acid bbls Gel bbls Foam bbls I Nitrogen Sand Quantity Ibs Open Flow Test Job Single Point Test Job Multi Point Test Job Fig85 Cost analysis module gr NMet Input a API Number Long Water i SumFluids 5 Prior kh Alter TestTy 12 345 1 1 255 12400 123 Open Flow 24 005 2 255 12400 123 Open Flow 24 005 255 13600 395 Open Flow 34 005 5 0 6000 3r8 Dpen Flaw 34 005 2 265 13650 100 Open Flow 24 005 2 255 9500 523 3 Open Flow 34 005 3 0 12100 262 34 005 234 11000 210 Open Flow 34 005 0 AA Open Flow 34 005 1374 Open Flow 34 005 289 Open Flow 34 005 5 793 Open Flow 34 005 134 Open Flow 34 005 120 Open Flow 34 005 141 Opern Flow 24 005 3 1210 Open Flow ee Oo Fig86 Inpurs that used to train the Neural Network I
13. Field Perforation Top ft fers Well Perforation Bottom ft fev Perforation Type Tubing Packer Shot Type um o Perforation Date 09 26 2005 Shots Per foot 2 MM DOA rr First perforation Previous Perforation Next Perforartion Last Perforation Fig49 Adding a complete new Well entering data for perforation 66 F Gas Storage Consortium JN Humber Well Count SED Adding New wel bere data Delete Find wellbore Stimulation well Test Adding Stimulation Well Mo Type Flush bbls Size of String 7 Water bbls Sand Quantity Ibs Stim From ft Acid bbls Sand Type Stim It Gel bbls Injection Rate bbls min of Shots Foam bbls Total Fluid bbls Fractured by N2 Summary Total bbls Date Alcohol bbls Break down Pr psi Cushion bbls ISIP psi SS Previous Stimualtion stimulation Last Stimualtion Fig50 Adding a complete new Well entering data for stimulation Gas Storage Consortium Number Well Count Adding New Well bore data petere Adding New Well Test Well Test Time hrs Field Pr pai Flow Pr psi Hate MctD First Reading Field Second Heading Well No Third Reading Test Date 09 26 2005 TestType Extended Time kh md ft NonDarcyCo eff 7 DeltaP psi 2 wel Test C3 PD Rate Skin s
14. LN teas 10 75 Fig4 Multiple Data Entries in Completion Table In the completion table the following notations used as casing description were replaced in place of different notations being used to have a standard definition NOTATIONS DISCARDED 18 Completion data was mostly checked for accuracy from the documents picture of which is shown below for a Well NEXT MENE r1 nlii S wetmesa Comp T Lesemg Reseed Recs EI Weste vane a _ Bom 3 green Components Dates a are e a lon so mm EIL L Downhole Components Dates lt 1990 are approximated un oa aee 00m Joo oes mem amen 3 393 1 831913 5 968 Tj Casing 256 81 3 1913 LL 39 99 18 4 1913 8 5 1913 ross 16 9 1968 631958 3000 41 4 Ancho 2857 2857 urs LE amica 16 9 1968 6 3 1968 3 000 44 4 Anchor 2857 2857 m 3i E EET Helps a TF 11 9 1995 1288 11 9 1995 amp s31
15. SRST Fig9 Stimulation data retrieved from a file 25 Each Microfiche contained dozens of documents from nrveices to valuable stimulation and well test data Microhflm scanners Each scanned image of the microfiche was searched and read for valuable data Pnnts were taken of the documents contammg data Waa Viale MI Wales Via per 7 Waleed Werde Habag Ca Wald l Maleate Verlag 5 342 1p 0 2 34205 10519 5 Ez Data was entered in the database Waleed D Dem Wiat Fig10 Microfiche to Database process 26 Following are pictures of some types of data formats for fracture jobs found in the records NITROGEN AMOUNT IN TOO BIFFERENT F 5 4 TEL LAN PLT wa cra ACID OR FRAC REPORT Fam T Dra 2 r POTE nimii n jens af Preti Horta PF E rig emo a ow DE rad c Tus Senne Dau Treriagea 7 50 90 Rash Fro Friar FL ins
16. Vertical Depth Formation 430 430 DATA 16 COMPLETION DATA Completion data mostly relates to the type and depth of casing liner tubing run in the gas storage wells The data type retained for the database includes the following API Number Field Name Well Name Well Completion Description Des Date Tubing Run Dt Tm Rn Outer Diameter OD Top of Casing Bottom of Casing Bot Casing Weight Weight 0 Casing Grade Grade t pede perc e Unfortunately the data was mostly in an excel file and had to be verified with well schematic drawings This led to the most unusual step in this project as it lead to reduction of valuable data available to us This was due to the erroneous and multiple data entry originally in the completion table Identification of the multiple entries and their removal from table was the most focused act of cleaning the data as omission of desirable records was unacceptable Following pictures show one of such flawed multiple data entries which were removed 17 pee IN ME Ter BL i1 54 005 31981 34006 91981 24006 91351 34 005 91981 Casing Casing A i 5 Seamless Dime Pipe 40 00 Surface Pipe 24 00 Casing 20 00 Casing FA Casing G20 Used Packer Eai ES Straight Anchor Packer Straight Anchor 2 irm ous CEO m Me m HET Joas i TETRIS Fr 25 3200 if T R F
17. When the Calculate button is pressed the program uses the permeability value k from the nearest well test and calculates skin The new value of skin can be saved in the database by selecting Save Results button Note Multi point test points give erroneous calculations if selected with other well tests as they are recorded one flow after another simultaneously not like Open Flow and single point tests which are recorded once a year If the selected well tests are not within 2 years then the following message will appear giving the user choice either to select other well tests or continue with the well tests selected 85 WELL TEST ANALYSIS gt gt 1 2 345 6 7890 22 The time between the well tests is mare than z wears Do you want to continue Build up test Analysis If any Multi point well test has a detailed data pressure and flow rate profile vs time then the Detailed MP Data column in front of that test will show YES It means that the data for this well test can be analyzed to estimate a value of permeability TestT yp 12 345 67890 5 11 1965 Open Flow 12 345 67890 5 12 1365 OpenFlow 12 345 67890 11 8 1965 OpenFlow 12345 57890 4 18 1955 OpenFlow 12 345 67890 4 24 1859 OpenFlow 1234567890 9 23 1963 OpenFlow Double click the well test in the Grid that pou want to select for Simplified Analysis or LIT Analysis Single click in the Detaled MP Data column to estima
18. apply it on all wells If the All Wells option is selected then the Select Well button will be enabled and the user can select the well the same way as shown in the previous section of the user Manual for Find a well r Type Well Number to find Start Typing Well Number 2 345 b B3 354 005 1051 All Wells 354 005 1052 34 005 1 0549 34 005 1 0552 34 005 10575 Salani fel 34 005 10657 Type of Analysis 34 05 1085 75 34 005 10795 34 005 1 0838 24 0085 11029 M Double click ta select There are 81 walls Cancel Fig83 Selecting a well for candidate selection process This section of module relates to the Optimization methods available User can optimize the stimulations according to only change in skin criteria or may choose to select the GA based on cost and skin option where he she can give different weight ages to cost and skin 96 Optimization Method 1 GA based on Skin Cost weightage 2 W eighbage Options Menu This software can cater for many varied situations These options can be selected from the options menu bar on the top of the form It contains following items Select controllable parameters Material cost Import NNet NNet Input values GA characteristic Export the Grid to Excel Select Well Test Type Following 15 a screen shot of the items in the Options menu tool bar Optimization Options Select controlable parameters Material
19. are also used by Columbia Trans Fig19 Gas Properties Simulator The slope from Pwf 2 vs time on semi log graph was used to find kh amp then skin The three values of skin were plotted on Q vs S graph and extrapolated to Q 0 to get True skin S 36 TRUE SKIN 2 00 1 50 1 00 0 50 1 00 Fig20 Calculation of True skin FLOW FLOW TENDED o mmo mem ommo Table 1 Draw down Test Results Shahab Mohaghegh January 2007 Razi Gaskari amp Kazim Malik Build up test In build up tests the slope drawn for Horner plot 15 after the time when well bore storage effects were found to be minimizing from previous draw down test This slope is then used to find the values of kh amp skin The True skin is found the similar way as in draw down test Fig21 Calculation of True skin Build up test Subcontract No 3040 WVRC DOE 1779 38 Final Report BUILD UP BUILD UP EXTENDED TEST1 TEST BUILD UP TEST 1 28E 02 1 28 02 Lm pom ne Table 2 Build up test results DRAW DOWN BUILD UP AVERAGE ACTUAL D 8202 1 510 1 165 0 0005 0 0012 0 0010 00126 Table 3 Average Results Due to large errors corresponding to estimating skin and kh values manually it was decided that for time being these values will not be entered in the database 39 Fo
20. data When the wells for which the user want the data to be retrieved have been selected click the Select Well Data button and select the parameters Table Names r Gas Storage Technologv 7 surtium WELL BORE DATA COMPLETION DATA PERFORATION DATA STIMULATION DATA WELL TEST DATA w Field ame Well LeaseM ame Classification Lat Section Township County State Operator TotaMerticaldepth Formation DtTmRun Bat Weight Grade Completion ype ate PerfT op PerfBattom ShotT ppe ShotPertt Select all well bore Select all Completion Select all Perforation Select all Stimulation lt lt Select Well Show Well Data SizeOfString amp StimFram F MoO Shots Fracturedby StimDate Water Acid Gel Foam N2 Alcohol Cushion Flush Sandljuantity TotalFluid Fig79 Selecting Well Parameters Click Show Well Data to retrieve the data w TestDate v TestType E v Time v FieldPress1 FlowingPress1 v Timez FieldPressz Field 2 Rates Times FieldPress3 FlowingPress3 Fates TimeE tended FieldPressE xt FlowingPressE xt HateE xt RESERVOIR DATA Porosity Temperature Gashravity Thickness Select all WellTest Select all Reservoir 93 r Gas Storage Technology Consortium Mumber 12 345 B7 830 12 345 B7 83 12 345 67690 12 345 B7 830 12 3
21. deliverability due to well damage with time A typical storage field has tens if not hundreds of production wells Each well will respond to remedial operations in its own unique way The well s response to the remedial operation is a function of a set of uncontrollable reservoir characteristics such as porosity and permeability and a set of controllable parameters such as completion and stimulation practices The objective of this project 15 to identify the combination of best candidate wells for the remedial operations that will result in the most successful program each year and consequently provides the highest return on investment The project deliverable is a Windows based software application that would perform the analysis and provide the list of wells and their corresponding remedial operation for each year based on the budget constraints identified by the user The state of the art 1n intelligent systems application that 1s currently being used extensively in the Wall Street 15 the methodology to achieve the objectives of this proposed project This methodology includes a hybrid form of artificial neural networks genetic algorithms and fuzzy logic Columbia Gas Transmission Corporation is the industry partner of this project and cooperated with the research and development team in order to ensure successful completion of the project The software application that 1s the deliverable of this project and is explained in much detail in this re
22. in the well To browse between different Completions First Completion Previous Completion Last Completion To move to the first completion previous completion next completion amp the last completion in the record click on the button assigned to it The completions are assorted in ascending order according to date tubing run for current well 60 PERFORATION wellbore Completion Perforation Stimulation Well Test Perforation 2 of 3 CAS i 2058 SeT Field Name Well Perforation 5 5 ep 1381 MMOD Ay Perforation Date jesni oo 2407 o ooo Perforation Top ft Perforation Bottom Ft Shot Type Shats Per foot Fig43 Perforation tab To browse between different Perforations Next Perforartion Last Perforation First perforation Previous Perforation To move to the first perforation previous perforation next perforation amp the last perforation in the record click on the button assigned to it The perforations are sorted in ascending order according to perforation date for current well 61 STIMULATION wellbore well No Size of String i15 Stim From t af Shots IN Dowellns Fractured by Stimulation 1 of 1 Type w ater N2 Water bbls Acid bbls 24 bbls ES Foam bbls N2 Mcf 0 Stimulation COMME eoo 040
23. rates are at Wellhead so we match the BHP amp THP with time 22 m e T ite LE THP psig BHP psig 2 ce er e time hrs Fig14 Bottom Hole Pressure profile for Multi Point test The multipoint test data 15 divided into Draw down amp build up test and each one 15 analyzed separately 39 Draw down test Analysis of drawdown tests was done as described above and following graphs were obtained 1 06 4 1 28E 06 1 26E 06 1 24E 06 1 22E 08 1 20E 06 1 18E 06 1 16E 08 1 14 06 4 1 12E 06 1 10E 06 0 0001 Fig15 Flow Test 1 Delta pressure squared vs time 1 06 8 DUE 05 OOE 05 Fig16 Flow Test 2 Delta pressure squared vs time 1 1 28 06 1 26E 06 1 24 06 1 22 06 1 20 06 1 18 06 1 16E 06 1 14E 06 1 12 06 1 10E 06 0 0001 0 001 Fig17 Extended Flow Test Delta pressure squared vs time Well bore Storage 1 172 log cycle half slope line Fig18 Log log graph For well tests after fracture half slope line is drawn and for un simulated wells unit slope line is drawn to find end of well bore effects and start of pseudo steady state Gas production Simulator was used to find the values of viscosity z factor and compressibility of storage gas at different pressure assuming Gas gravity 0 585 amp Temperature 75 F 535R that
24. 17 Fig76 Selecting Ohio County Selecting wells by stimulation year The user can select the option button for stimulated year and input the year values If Select Wells button is clicked then all the wells that have been stimulated between these years will be shown in the selected wells list box 91 Gas Storage Consortium MED a ER cux SELECT WELL DATA SELECT SELECTED WELLS SELECTION OPTION STATE COUNTY API Well Number amp Year of Stimulation 34 139 20515 1333 O 34 005 22979 1334 f Stimulated between year 1932 and 4996 C Em Off set Wells Fig77 Selecting wells according to stimulation year Selecting offsets wells form a well The user selects the offset option and the well near which he wants to find the off set wells and then enters the distance of off set in kilometers If Select Wells button 1s clicked then all the wells that are off set of the selected well will be shown in the selected wells list box F Gas Storage Consortium T i 1 E d X SELECT WELL DATA SELECT SELECTED WELLS SELECTION OPTION STATE COUNTY Well Number 34 005 01272 24 005 02420 1 34 005 02987 34 05 0283 34 D05 023101 34 005 02960 34 005 35833 34 005 35225 34 005 37 782 Stimulated between year and 2 Em Off set Wells lt lt Main Menu Select Well Data Fig78 Offset wells 92 Display the selected wells
25. 2 Casing 1291 920 Fig5 Well bore data retrieved from a file 19 Analysis raw data vs refined data Please note that multiple data entry was the major reason for the reduction in the refined data from the initial data DATA FIELDS Run u E e 3 INITHAL DATA AVAILABLE FINAL DATA AVAILABLE 500 3000 2500 2000 1500 1000 500 COMPLETION DATA API Number Field Name Des cription Date Tubing Run Top Casing 2909 2909 2909 1781 1607 COMPLETION DATA INITIAL DATA AVAILABLE m FINAL DATA Fig6 Data addition and refinement for Completion Data 20 PERFORATION DATA This data set contains all the information relating to the perforations done on the gas storage well like perforation top amp bottom depth and shots per foot Following are the data types included in this type of data set Well API Number Field Name Well Name Completion Type Perforation Date Perf Date Perforation Top Perf Top Perforation Bottom Perf Btm Shot Type Shot Per foot Shot Per ft ee The picture of a document showing this information is shown below 21 7 5 Fan bees A Tes apu ed senhe bry ta feel A71 Rumba Goud Cate Acreage ann as o mm _ a _ TEM EN Oper M aee chm Fipa
26. 45 B7 880 12 345 B7 83 12 345 B7 830 Field ame 20 62222 aU 52222 2D 52222 20 62222 20 62222 20 62222 20 62222 B2 27472 Bdr B22 472 B22 472 B22 472 B22 4 2 B227472 TestDate 9 23 1353 4 18 1355 11 8 1355 8 12 1355 5 11 1355 41 24 1353 2 12 2004 TestType Open Flow Open Flow Open Flow Open Flow Open Flow Open Flow Multi Point SELECTED WELL DATA FieldPress1 FlowingPres Aatel Select Parameters Export to Excel Fig80 Result of the wells amp parameters selected 94 CANDIDATE SELECTION This module will appear on selecting the Candidate Selection button from Main Menu Y Gas Storage Technology Consortium File Help MAIN MENU T 1 E in e s E 5 j Y an Ee ul Edit Well Data View Well Data Canditate Selection West Virginia University 7 28 2006 8 03 AM Fig81 Start Candidate Selection form main screen For intelligent candidate selection of wells it is very important that only valid data is given to the Neural Network NN for training Valid data is one which will not degrade the performance of the NN and is useful in NN training Optimization Options API Number FieldName Well Lease ame Llassificatior Lat Long i 1234 Az SPECIAL 52222 27472 34005 10517 10517 L amp E JENF ACTIVE 705533 205 34005 10527 10527 E M DAHL ACTIVE 205033 4 34 005 10549
27. 508 1 4872 27 814 1025 163 48 115 02 7866 93 557 4 797 387383 5144 27208 11 89 43459 35648 718 6 2864971 1374 5 782 1598780 0 859 27 1135 132 83 22757 14858 1115 46 208 5042 496186 1 0253 108 245 18332 7 24701 20427 2407 05 793 505 699456 0 7801 257 211 28 7 19515 220 95 1819 96 134 5049 774901 1 5019 378 218 08 41 05 235 04 5107 63 120 5049 793018 1 654 701 21128 10946 228 07 2172 21 5 39 770906 1 0117 11 88 446 62 364 4 562 87 29647 75 5699 1637575 26374 956 144 34 177 88 15432 5999 26 5043 536900 1 4835 8 182 52 223 25 194 63 1457 71 5 026 Bragd 1 1274 1133 44h 71 3629F R3 PANEN AF AAP 1 8092 Export result to Excel ll lo ll lol ll cl Fig91 Optimization result for selected wells The wells are ranked according to change in skin as shown in the figure below These ranked wells and the optimized stimulation data now can be exported to excel by selecting Export to Excel in the Option menu of Candidate Selection module Wells Ranked according to Change in Skin Candidate Selection Options Water Acid Gel Foam Ne SumFluids SandQuai After T estT yy G After Skir G Cost 3 03 243 7 1 67 26 65 24841 3698 63 957 5 049 88287 1 35 5 94 164 21 17 234 77 1731 6810 18 105 5 046 610748 103 97 11 89 446 62 0 4 3544 562 87 29647 75 1210 5 699 1637579 2 03 1 65 192 45 1 69 33343 18782 176 13 308 7 5 048 713734 204 5
28. 7 11 87 471 2 019 34273 687 62 29941 29 1712 5 878 172732 5 75 11 34 121 85 1 63 25277 141 57 469 67 123 5 024 458092 142 45 11 75 507 91 34712 662 09 28356 16 1880 5 894 1855768 7 43 gab 222 79 1 69 7202 241 16 410 96 576 5 048 80341 6 52 23481 1 15051 243 04 1457 71 100 5 047 85543 7 43 2 69 212 33 64 98 22409 3992 17 100 5 027 774088 180 26 11 33 44671 362 56 540 83 29060 67 1770 5 872 1646979 4 05 11 21 103 03 0 03 46 09 118 32 23835 62 100 4L 457 38385 261 28 11 85 456 03 1 1 9536 730 27 26125 24 1317 5 747 167386 3 38 882 172 06 235 38 1883 2289 63 117 5 64007 1 35 3 71 205 53 27222 212 26 0 122 5 024 75735 11 48 378 218 08 41 05 23504 5107 53 120 6 049 793018 3 38 2 44 265 15 141 87 27253 117 42 209 5 045 963812 n nas 234 R4 124001 1 lol ee a ee a a 2 cl c c 1 Fig92 Rank the optimization result based on delta skin in order to find the best candidates 101 CONCLUSION The main aim of this study was to find the re stimulation candidate wells with the given data without trying to spend thousands of dollars on well test and gas reservoir simulators Detailed analysis of well tests performed on the storage field was done and intelligent tools like Neural networks to predict the Skin and Genetic Algorithms were used to optimize the stimulation and to select the best stimulations for a well The following conclusions can be dra
29. Gas Storage Field Deliverability Enhancement and Maintenance An Intelligent Portfolio Management Approach Final Report Reporting Start Date September 1 2004 Reporting End Date December 31 2006 Report prepared by Shahab D Mohaghegh Ph D Principal Investigator Razi Gaskari Ph D Co Principal Investigator And Mr Kazim Malik Petroleum amp Natural Gas Engineering West Virginia University Morgantown WV 26506 Telephone 304 293 7682 Fax 304 293 5708 Reporting Issue Date January 2007 Subcontract No 3040 WVRC DOE 1779 Report prepared for GSCT Consortium Director PSU Energy Institute The Pennsylvania State University C211 Coal Utilization Laboratory University Park PA 16802 2309 Telephone 814 865 0531 Fax 814 685 3248 Email jlm9 psu edu DISCLAIMER This report was prepared as an account of work sponsored by an agency of the United States Government Neither the United States Government nor any agency thereof nor any of their employees makes any warranty expressed or implied or assumes no legal liability or responsibility for the accuracy completeness or usefulness of any information apparatus product or process disclosed or represents that its use would not infringe privately owned rights Reference herein to any specific commercial product process or service by trade name trademark manufacturer or otherwise does not necessarily constitute or imply its endorsement recommendation or favoring
30. IVE ACTIVE Nu e e e e 40 6961 19812 44 006 60255 SPECIAL 40 7133 103921 34005 83821 ACTIVE 40 0383 13926 94 006 80285 ACTIVE 40 7 105 103929 34 005 93929 ACTIVE 40 7113 3931 34 005 93931 40 6889 ACTIVE Corr ected Number Active ACTIVE ACTIVE ACTIVE ACTIVE ACTIVE ACTIVE ACTIVE ACTIVE ACTIME ACTIVE Fig2 Correction of Wrong API number in data 40 5703 40 6735 40 6767 40 6229 40 6983 40 6739 40 6797 40 6835 40 6605 40 6835 40 7378 40 6855 82305817 822471 32 2719 34 B2 2358 12 B2 2611 10 82 3047 7 B2 347B 24 B82247 5 B2 2375 25 82 3017 17 B2 3338 19 82 3453 54 H2 3417 24 82 2905 9 82 3061 B 3375 19 82 3400 24 82 3484 13 B2 3355 19 82 34513 B2 2231 56 A2 3536 13 E 22 9330 719 0 3453 24 40 3417 24 42 2925 8 amp 30618 0 337548 EQ 3408 24 482 3494 13 52 3356 19 82 345 13 40 223 56 2 3538 13 82 2514 10 40 6703 40 6736 40 5761 40 6229 40 6353 40 5739 ers 40 68 5 40 40 6850 40 7378 40 5855 40 6953 GREEN ASHLAND GREEN ASHLAND VERMILLION ASHLAND GREEN ASHLAND GREEN ASHLAND GHEEN ASHLAND MONROE RICHLAND GREEN ASHLAND VERMILLIION ASHLAND GREEN ASHLAND GREEN ASHLAND MONROE RICHLAND MONROE RICHLAND HANOVER ASHLAND GREEN ASHLAND GREEN ASHLAND MONROE RICHLAND MONROE RICHLAND GREEN ASHLAND MONROE RICHLAND VERMILLION ASHLAND MONROE RICHLAND GREEN GREEN
31. ailed MP Data column to eshmete permeability For the weites where ES is vaithen Fig65 Well Test Analysis Module 1 This section contains all the data in a grid form API Number Date of well test Test Type kh value Skin value Peak Day rate Absolute Open Flow and information in YES or NO form if the Detailed Multi point data Pressure profile amp flow rate vs 80 time is available for a given test or not The back color of selected well test is yellow while of stimulation is purple The first well test 1s selected by default 2 This section contains instructions as how to select well tests for analysis Single click on any well test will make it the current well test with background changed to yellow and by double click it will be selected for Simplified and LIT Analysis If the Detailed MP Data for a well test 1s given then it can be selected for permeability analysis build up test by single click on the cell where YES 15 written This way the build up test analysis module will show up 3 This section shows the time of different well tests which are indicated by three types of markers and stimulations on a well which are represented by straight blue vertical lines The selected well have the similar marker according to its test type but its color is dark green Once any well test is double clicked it is selected and added in the list box of simplified and LIT Analysis 81 S
32. at a time so if there is more than one well test for which the value of has to be assumed then they should be selected one by one F WELL TEST ANALYSIS gt gt 12 345 67890 x E Simplified Analysis OS Selected Well Tests T B Eu H 12 12345 57890 5 12 1965 714 12345 67890 4 18 1966 OpenFlow 15 12 345 67890 4 24 1969 Open Flow 9 29 1969 OpenFlow pen Flow log Delta P sqr Fig69 Simplified well test with one well test before or after stimulation 84 Laminar Inertial Turbulent LIT test Analysis The screen shot of the LIT analysis 1s below r WELL TEST ANALYSIS gt gt 12 345 67890 m LIT Analysis vi 2 12 345 5 890 5 12 1965 OpenFlow E 3 57990 rai open EE E E m m C gt log Rate Clear gt gt Parameters Calculatec 0 0252621877 Fig7 0 LIT well test analysis 1 This section is the same as for Simplified Analysis 2 In this section the well test points are drawn on log log plot of Flow rate MMcfD vs Delta pseudo pressure bq There is no need to draw a slope line in this plot Instead the points can be selected by visual inspection that they form a straight line and that they were conducted preferably within 2 years In the snapshot above well test points 2 3 and 5 have been selected to calculate Skin 3
33. by the United States Government or any agency thereof The views and opinions of authors expressed herein do not necessarily state or reflect those of the United States Government or any agency thereof ABSTRACT Portfolio management a common practice in the financial market 15 essentially an optimization problem that attempts to increase return on investment The objective of this project 15 to apply the state of the art in optimum portfolio management to the gas storage field in order to optimize the return on investment associated with well remedial operations Each year gas storage operators spend hundreds of thousands of dollars on workovers re completions and re stimulations of storage wells in order to battle the decline in deliverability due to well damage with time A typical storage field has tens 1f not hundreds of production wells Each well will respond to remedial operations in its own unique way The well s response to the remedial operation 15 a function of a set of uncontrollable reservoir characteristics such as porosity and permeability and a set of controllable parameters such as completion and stimulation practices The objective of this project 15 to identify the combination of best candidate wells for the remedial operations that will result in the most successful program each year and consequently provides the highest return on investment The project deliverable is a Windows based software application that would perform
34. e keeping your well bore tab as active Editing only completion perforation stimulation well test data 1 Click on Delete button ite keeping that tab active of which you want to delete the data except well bore tab 74 Shahab Mohaghegh January 2007 Razi Gaskari amp Kazim Malik m mese XT Em Main Menu First Well Previous Well Mex Well Last Well Completion First perforation Previous Perl Fig57 Deleting perforation record You will be greeted with the above message to make sure that delete button 15 not accidentally pressed 2 Click on yes uem if you want the selected record to be deleted Subcontract No 3040 WVRC DOE 1779 75 Final Report Undo the edit add operation To undo the edit or add operation before they can be saved click undo button Finding a well Follow the following procedure to find a well for which you have some idea of its API well number Find Click on Find button The following screen is displayed Select Find button td Type Well Mumber to find Sart Typing Well Humber 34 005 02420 F Gas Sterape Techmolepy Comertivn Vell Count well E Z of 43 AP Welle 3150505126 Duas wel Na nnm Lease Mame ETMA LAND Fig58 Finding a well 76 WELL TEST ANALYSIS To perform well test analysis on a well and draw graph of Peak day rate and Absolute open flow use the option command buttons below
35. e package it 1s recommended that the storage filed have a minimum of 75 wells wells with data that can be used for analysis TABLE OF CONTENTS i ui Se Table OEC everett ISI couse IN pe EE US I sil DO MERO LS OP IS cic a aye ES REDIMI cited oyu inate outdo SUE Execu ve Results amp Discussions 2 emere es CONCIUSIONS eee Ore ee ee ee ee eee ee LIST OF FIGURES Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Figure 6 Figure 7 Figure 8 Figure 9 Figure 10 Figure 11 Figure 12 Figure 13 Figure 14 Figure 15 Figure 16 Figure 17 Figure 18 Figure 19 Figure 20 Figure 21 Figure 22 Figure 23 Figure 24 Figure 25 Figure 26 Figure 27 Figure 28 Figure 29 Figure 30 Figure 31 Figure 32 Figure 33 Figure 34 Figure 35 Figure 36 Figure 37 Figure 38 Figure 39 Figure 40 Figure 41 Figure 42 Figure 43 Figure 44 Figure 45 Figure 46 Figure 47 Figure 48 Well bore data retrieved from a file Correction of Wrong API number in data Data addition and refinement for Well bore Data Multiple Data Entries in Completion Table Well bote data retrieved from a file isses nda Ea a RR Dn Data addition and refinement for Completion Data Perforation data retrieved from a file Data addition and refinement fo
36. epresenting the table in the database of the software O O O O O OQ Well bore Data Completion Data Stimulation Data Perforation Data Stimulation Data Well Test Data Reservoir Characteristic l These are the fields of the table Each field represents one characteristic of the table and each row 15 one record If the user 15 not clear about any field then he she can drag the screen cursor to that field name and the comment will appear like in the picture below where it will give a little explanation its format and an example so that the user understands what sort of data to enter in each field Fig34 Comments that shows format of some cells in Template Excel file 53 2 This section has two sets of warnings for the user entering data One 15 to not edit or change number of Titles in all the worksheets or worksheet names and the other is to add only unique Number in worksheet bore Data and all dates in worksheets where required This has been done as the data is retrieved from the template according to some specific format and non presence of any data in elementary field might stop program from using that record All the elementary fields background is orange red while others are in green 3 This section shows all the worksheets in the Template file Import Data from filled out Template Gas Storage lechnology Consortium Fig35 Import data from filled out Template If this option is selected f
37. evious WellTest Mex WellTest Last Wellkest To move to the first well test previous well test next well test amp the last well test in the record click on the button assigned to it The well tests are sorted in ascending order according to well test date for current well Adding a new data One can add a complete new well or just only a new well bore completion perforation stimulation well test data by following method Adding a complete new well data 1 Click on the Add New button add new while keeping your well bore tab as active 63 Gas Storage Consortium Humber 324 005 601 83 Adding Mew w ell bore data Main Menu pete em Well Court YW ellbore Completion Stimulation Well Test well umm Field Mame nasse welNo Lease Mame Classification Latitude Longtitude Operator Section Township County State Fig46 Adding a complete new Well well bore tab The following messages will pop up If you want to add the complete new well bore data then click No button is If you don t have the dates of Stimulation Completion Perforation amp Well Test data then click Yes 5 and then add them one by one Gas Consortium vou want bo Add only Wellbore Data Otherwise vou must know Ehe dates of Stimulation completion Perforation amp Well Test Follow
38. ew Well entering data for stimulation Adding a complete new Well entering data for well test Result of adding a complete new well 00 ccc eeeeeeeeeeeeeeeeees Editing completion data Saving completion data Saved completion data Deleting perforation record Well test Analysis Option in well test tab Show Chart Peak Day Rate Show Chart Absolute Open Flow SHOW CNIS SKIN neriie o Show Chart All Well Tests sares E doti n daa Well Test Analysis button Well Test Analysis module Well Test Analysis tool nmm Draw a line and calculate the slope Simplified well test analysis tool Lecce cece cece cece eee ees Simplified well test with one well test before or after stimulation LIT Well test analy SiS tei Multi point well test analysis 0 0 ccc cc eeeee eee e eee Well extended pressure profile Selecting the build up selection from pressure profile Diagnostic plotanalysis dUpfe tede tM eT Res Calculating skin from Hornet plot Selecting OU B REO ns Selecting wells according to stimulation year Borse EE Selecting WelllPataietels Result of wells amp parameters selected Start
39. f Template file Comment that shows format of some cells in Template Excel file Import data from filled out Template Remove all data from Exit from file REVERSE ES Help menu PERSE about screen from help menu Browsing through the well bore data Well bore tab scisrsenoeanteneaateriscdere PESER NES PECES Completion TAD ho SURE st na dois PEMOGAUOMMAD GdimulgtiOH flr m Adding a complete new Well well bore tab Adding a complete new Well completion tab Adding a complete new Well entering data for wellbore Figure 49 Figure 50 Figure 51 Figure 52 Figure 53 Figure 54 Figure 55 Figure 56 Figure 57 Figure 58 Figure 59 Figure 60 Figure 61 Figure 62 Figure 63 Figure 64 Figure 65 Figure 66 Figure 67 Figure 68 Figure 69 Figure 70 Figure 71 Figure 72 Figure 73 Figure 74 Figure 75 Figure 76 Figure 77 Figure 78 Figure 79 Figure 80 Figure 81 Figure 82 Figure 83 Figure 84 Figure 85 Figure 86 Figure 87 Figure 88 Figure 89 Figure 90 Figure 91 Figure 92 Adding a complete new Well entering data for perforation Adding a complete n
40. f information is the User Manual for this software and second 15 the Formulas used in this software and third About form which shows the system information and software contributors C West Virginia University Petroleum amp Natural Gas Version 1 0 GAS STORAGE TECHNOLOGY CONSORTIUM Project Gas Storage Field Deliverability Enhancement Product Gas Storage Wells information Database and Candidate Selection Funded By U S Department af Energy Contributors kazim Razi Gaskan Shahab D Mohaghegh Copyright West Virginia University 2006 Fig39 about screen form help menu 56 Edit View Well Data This screen has all the well data in the form of five tabs for five database tables that can be edited viewed or a Well Test Analysis can be performed in the Well Test tab Gas Storage Consortium Previous Well I RESUME Humber 34 D 5 333 33 W ellbore API Well No 34005 33338 FieldName LUCAS En Lease Marne MAURE Classification AUTE ACTIVE Well Well 33 of 431 Last Well Well Count Latitude 40 74306 Longtitude 222375 Section 5 Township VERMILLION County ASHLAND State oH Operator Columbia Gaz T D ft 2879 Formation Clinton Fig40 Browsing through the well bore data To browse between different wells Previous Well Last Well Completion Stimulation We
41. f the user wants to look at the Neural Network inputs being used then NNet Input option will take the user to a new form as shown above where all the inputs are shown Keep in mind that this grid can only be seen once and that only after the Genetic optimization has been applied If some material is not available for stimulation then still the user can optimize the stimulation by de selecting that material from the Select controllable parameters option The materials not enabled are the ones that are not being used by the Neural Network use 98 Select controlable parameters according to availability E W Foam Water W Ne W Acid W Gel n Iw Sandiuantity Close Fig87 Select the controllable parameters in optimization process A new Neural Network can be used if the data is changed or appended by importing its ida file When a new Neural Network is imported it might change the optimum GA parameters The user can change them from GA characteristic option The default values are always loaded at startup as shown in figure below but can be changed by user GA Charaterictics Population Size 500 Cross over Rate mm Mutation Rate of Generations Close Fig88 Setup GA pentameters If one of the Neural Net inputs is well test before stimulation the type of the wells test in optimization process should specify here Well Test Type OpenFlow Single Poirt C M
42. hat tab active for which you want to edit the data except well bore tab EST Main M TTT TTT TTT Add New Save undo Humber Well Count 12 345 67892 Well tt 433 of 433 Edit wellbore Completion Stimulation Well Test Editing Completion Data Field OD in F Gas Storage Consortium Well Na ft Ft Description weight Date Tubing Hun MMOD Y sees First Completion Previous Completion Next Completion Last Completion Fig55 Saving completion data The background color of all text boxes of that tab will be yellow indicating that they are ready for entering data Enter the data The dates for completion perforation stimulation amp well test job should be known Click the Save button _save 73 Result of editing only completion data has Storage Consortium d E Number Well Count virna Find Completion 1 of 1 Field LU CAS OD in ft Well 67892 Description SurfaceCasing Date Tubing Run 8 26 2005 MM DD Bottom ft Weight Grade First Completion Previous Completion ext Completion Last Completion Fig56 Saved completion data Deleting data One can delete complete well or just only delete completion perforation stimulation well test data by following methods Deleting a complete well data 1 Click on Delete button whil
43. he database for this software consists of five main tables Well bore Data Completion Data Perforation Data Stimulation Data Well Test Data Reservoir Characteristic Data purae so The number of a well 15 the primary key in this database so it must be known before adding a record and cannot be duplicated GasStorage Database Access 2000 file format 25 Open B Design Mew X ES Objects Create table in Design view Create table by using wizard Queries Create table by entering data Completion Data F MaterialCost_Daka Reports Perf Data Pages Reservoir Data Stimulation Data Wellbore Data WellTest Data Macros Modules EJ EJ EJ EJ EJ EJ EJ k 021 IS Groups Favorites Fig30 Screen shot of database showing different tables 50 The software starts with the main menu screen with six options Complete list of items and sub items in the above command buttons is shown below File Help EDJ AM Fig31 Main Screen of software o Create Template o Import Data from filled out Template o Remove all data from database o Exit o User Manual o Formulas o About Edit Well Data o Well bore o Completion o Stimulation o Perforation o Stimulation o Well Test Well Test Analysis Tool o Reservoir o Find a Well View Well Data 51 o Select State amp cou
44. ia Gas Transmission Corporation 15 the industry partner in this project and provided us with very valuable data and in depth knowledge about their gas storage field operations The data in very crude form was provided to the research and development team in the last week of March 2005 The team extracted valuable data and organized it in a form of database with generic make up in order to be reusable Windows based software was developed which can help the user in viewing and later populating the data with easy to use interface One of its modules provides the user with all the valid stimulations required as an input for Neural Network A Neural Network was trained in order to predict skin for different stimulation parameters A Genetic Optimization tool was developed and associated with the trained Neural Network in order to find the optimum stimulation parameters The software ranks the well according to maximum change in skin value or and stimulation cost for well Then a decision 15 made to stimulate a well or not accordingly 11 DATA MADE AVAILABLE AND ITS FORMAT The research and development R amp D team was initially provided data in MS excel worksheets On further request some pdf files with well schematics well test files and well summary files were provided but still the required data especially relating to stimulations and well tests was so scarce that the team in July 2005 went to the Columbia Transmission Corporation Off
45. ice in Charleston WV to get more information Retrieval of data from different files and thousands of microfiche was taking so long at the office that 1t was decided that West Virginia University lab facilities will be used to read thousands of microfiche So for the next few weeks the team concentrated its efforts on data collection That data could be segregated into five main tables each relating to specific characteristic features of the gas storage wells The five characteristic features are as below 1 Well bore data 2 Completion Data 3 Perforation Data 4 Stimulation Data 5 Well Test Data 6 Reservoir Characteristic Data 12 WELL BORE DATA It includes basic features of the well like location depth well name etc Data about well bore was retrieved mostly from well schematics and well summary reports The data already provided by Columbia Transmission Corporation was also verified The complete list of the data type retrieved 15 as below API Number Field Name Well Lease Name Classification Latitude Lat Longitude Long Section 9 Township 10 County 11 State 12 Operator 13 Total Vertical Depth 14 Formation VR aM Picture of of the forms from which this data was retrieved 15 on next page 13 MTITI Ea lI xs Tes 14e SE angu uve
46. implified Analysis The screen shot of Simplified Analysis tab is below with Well test 2 to 7 selected for analysis 6 12 1965 Open F low 11 8 1355 Open Flow 71234567896 8 2 Flow 7 12 345 67890 7 12 2004 Mult Powt log Delta P sqr log Rate Well Deliverability Parameters Calculated from above value of n Selected Well Test value PDrate McID AUF Calculate C PDR AOF W Open Flow 5 Single Point Moulti Poin Selected Test Stimulationi Type Fig66 Well Test analysis tool 1 This section contains the list box which has the entire well tests selected for an analysis Any test now again can be selected or de selected by using the check box in front of it Once the well tests are selected then they can be drawn on log log graph of Flow rate McfD vs Delta Pressure Square Delta P sqr by selecting the Draw Points button This graph can be cleared by selecting the Clear gt gt button also if the well tests drawn need to be changed 2 Once the data points have been drawn the user can draw a line in the picture box keeping left mouse button held down like shown below 82 nplified Analysis LIT j log Dielta P sor ur log Rate Slope of line m 127220432 Fig67 Draw a line and calculate the slope Select Calculate n button to find the slope of the drawn line The slope will be calculated in front of n tex
47. ing screen appears if No 15 clicked 64 Gas Storage Consortium EE nm wc D DESI z Main M ERE ETT Main Number Well Count Adding New Well bore data Delete find wellbore Well Test Adding Mew Completion Field Mame OD in Tene well No ap r Ft Description Weight Date Tubing Run Riese First Completion Previous Completion Next Completion Last Completion Fig47 Adding a complete new Well completion tab The background color of text boxes of all tabs including well bore tab will be yellow indicating that they are ready for entering data 2 Enter the data in all the tabs The dates for completion perforation stimulation amp well test job should be known 65 t Gas Storage Consortium Main M Number Well Count Adding New el bore data Delete Find Lade API well 12 345 67881 Longtitude 822527 perator rs Field Mame lucas Sesion 89 Township GREEN Lease Mame ASHLAND Formation Cinto Classification ACTWE ss p Fig48 Adding a complete new Well entering data for wellbore er Gas Storage Consortium Number Well Count 234 005 601 83 Adding Mew Wrell bore data Delete Adding Hew Perforation
48. ining of the neural net is shown in Figure 25 44 E GRNN Network GSTIC May 9 m Scotter Plot cec PIE Erer Plot Frag Error Piel fra 5 Draw Option Marker Error Places Fig25 Accuracy of training data for the Neural Net The calibration and verification of the Neural net is shown in Figure 26 and Figure 27 respectively After the accurate results of this GRNN the software was updated to use the GRNN generated files to be used 1n the Genetic algorithm 9 Fig26 Accuracy of calibration data for the Neural net 45 Shahab Mohaghegh January 2007 Razi Gaskari amp Kazim Malik Fig27 Accuracy of verification data for the Neural net Subcontract No 3040 WVRC DOE 1779 46 Final Report GENETIC OPTIMIZATION MODEL Genetic Algorithm was written to optimize the stimulation parameters used in the neural net Out of the 11 input parameters 7 can be varied to obtain optimum skin The range of these variables was calculated and accuracy desired was determined to design the length of the chromosome of Genetic Algorithm GA that will be required The calculation is shown in the table 4 for the chromosome length if all the parameters are selected GA Accuract Lic curacy tange Thr mem me im i Eerie redii ht E
49. ir Temperature Gas Specific Gravity Reservoir Thickness ON c 43 NEURAL NETWORK MODULE The Neural nets are very powerful in predicting non linear relationships As the relationship between skin and stimulation parameters 1s non linear and very complicated thus neural nets are used which are very good at it With skin values before and after the stimulation calculated and stimulation parameters known we can now use these valid stimulations to train the Neural Network to use it as a prediction tool Intelligent Data Evaluation and Artificial Network IDEA software by Intelligent Solutions Inc was used to design the neural network This software 15 very versatile in making different nets with different training algorithms Generalized Regression Neural Net GRNN was used to train the neural net The net had 11 inputs and 1 output as skin The source of data for the neural net is given in Figure 24 Inputs Source Sum Fluids cum of iter Water bbls GA Acid bbls Gel bbls Foam bbls I Sand Quantity lbs 11 After T est Type 3 IVIulti Point 2 Sing e Pumnt l Open Flaw Fig24 Neural Network Inputs and their source Out of the 78 valid stimulations available the Neural net was trained on 60 data items while 14 were used as calibration data and 4 as verification data The Neural network showed very good results for all three types of data The screen shot taken from the IDEA software for tra
50. know Ehe dates of Stimulation Completion Perforation amp Well Test Select accordingly Gas Storage Consortium Save Undo Number Well Count 12 345 6 7892 Editing wellbore Data Edt Delete fna Wellbore Completion Stimulation Well Test Latitude 40 7366 API Well Ma 12 345 67892 Longtitude 222527 Operator TNS Field uca Section welNo SER Township Lease KAM Classification ASHLAND Fig53 Editing well data 71 has Storage Consortium Main Menu xe esee eS TUTE EEE LE LE LLL LEE ELL AddNew save Undo Number Well Count 7 Eding Welbore Data Edt Delete find Wellbore Stimulation Well Test Editing Completion Data Field M ame OD in Tes well No tt Bottom ft Description W eight 6 Date Tubing Run MM DD AY YY S First Completion Previous Completion Next Completion Last Completion Fig54 Editing completion data The background color of text boxes of all tabs including well bore tab will be yellow indicating that they are ready for entering data 2 Enter the data in all the active tabs The dates for completion perforation stimulation amp well test job should be known 3 Click the Save button 72 Editing only completion perforation stimulation well test data 1 Click on the Edit button while keeping t
51. ll Test To move to the first well previous well next well amp the last well in the record click on the button assigned to it The records are sorted ascending order according to well number API Number amp Well Count BB E Humber well Count 34 005 93939 well 39 of 431 57 The progress bar shows the relative position of the record and well count shows the current well position in the well bore database out of the total records The API number of the current well 15 also displayed Takes you back to the very first screen of the program Back to main menu Editing Tools eae eere These buttons will help you to add new record edit or delete it or find well for which you want the data to be retrieved if you know its API number 58 Shahab Mohaghegh Razi Gaskari amp Kazim Malik Different Tabs WELL BORE 34 15 3 3333 LL AS das LTIVE January 2007 Columbia Gas Fig41 Well bore tab This tab contains all the data pertaining to the name location amp some main features of the current well Subcontract No 3040 WVRC DOE 1779 Final Report 59 COMPLETION Completion 4 of 6 Field Name Lu Cas OD in Top ft well 326 Bottom ft Description Casing Weight 17 00 Date Tubing Run 1 2 1963 SMLS USED Fig42 Completion tab This tab contains all the data relating to different completion run
52. llowing are some pictures of the documents to show the different format in which the data was presented in files and microfiche ___ ____ idii EM r3 Obs PLF Sted whee dora liebe albas era Hr Ws o LL o addo im M miit __ FA LM eL eus Sires 1 wl wi ET IM il Meare figir amd Pasa Ukncg an Shaded 5 bub Hirvis ai TE Pear ga eto WELL ur ge WELL i SS uw cmm pho emm lt a os 5 16 Minutes 15 Hinutesa FLOW RATE 1 Fig22 Retrieving flow rate of an open flow test 40 Laminar Inertial Turbulent LIT Test Analysis of data from isochronal type test using Laminar Inertial Turbulent LIT flow equation will yield considerable data This method can also be used to find skin of a well from single point test when the value of permeability of reservoir is known from prior multi point test The LIT equation 15 written as Ay V p TY wr sc bq Pressure drop due to Pressure drop due to laminar flow and well conditions Procedure for calculating Skin from LIT analysis for known permeability value 15 a
53. nty o Select Wells o Selection Options o Select Well Data Candidate Selection File The file menu can be accessed from the top left corner of menu bar It contains four options o Create Template Import Data from filled out Template Remove all data from database Exit T Gas Storage Technology Consortium WI Help Create Template Import Data From Filled out Template Remove all data From database Edit Well Data View WellData Canditate Selection NG West Virginia University of 15 2006 2 56 Fig32 File Menu options Create Template By executing this option first the user need to select a location in hard drive in order to save Template file Once the Template is successfully created in the hard drive a message will appear indicating the user that the template file has been created Following is the screen shot of the Template file showing the Well bore data 52 SP Ey Sa ae Ty ETE i i L B TITTITITITITITIITIITIITIITITITITPIEITDITIITIITITEI acies esee donna dadda PA PA vT tes Please enter only unique Number in worksheet Yielbore iw ae 57 Wm d 999 P EE Fig33 Screen shot of Template file It has six worksheets each r
54. omp sar 2823 o 1 Fig7 Perforation data retrieved from a file 22 Analysis of raw data vs refined data PERFORATION DATA DATA FIELDS AVAILABLE 392 Sf 57 51 48 392 392 392 209 383 DATA AVAILABLE 392 5 Sf 51 48 392 392 392 209 383 PERFORATION DATA Field Name Well Perf Date Pert Bottom Shot Type Shot Per ft API Number Completion Type 9 t EE m Number of Data Entries OINITAL DATA AVAILABLE B FINAL DATA Fig8 Data addition and refinement for Perforation Data 23 STIMULATION DATA Stimulation data is one of the most significant datasets about the storage wells Because of this it was very important that we have maximum records of valid stimulations Following data type is used to represent stimulation 1 Well Number 2 Well Name 3 Size of String 4 Stimulation From 5 Stimulation To 6 No Of Shots is Fractured by Stimulation 9 Stimulation Date 10 Water 11 Acid 12 Gel 13 Foam 14 Nitrogen 15 Alcohol 16 Cushion 17 Flush 18 sand Quantity 19 sand Type 20 Injection Rate 2 Total Fluid 22 Breakdown Pressure 23 Unfortunately initially we didn t have much data about the stimulations being done in this Lucas field With this in mind every record with Columbia Transmission Corporation was carefully examined The la
55. on Value O PbRae MeD CAF CValue First WellTest Previous WellTest Next WellTest Last Welltest Fig51 Adding a complete new Well entering data for well test 67 3 Click the Save button Sae Result of adding of complete well data Gas Storage Consortium save Number Well Count 124 345 6789 Well 433 of 433 idit Delete Find Latitude 40 7366 API wel 12 345 57892 Field Mame Lu CAS Longtitude 82 2527 Operator CTHS Section 33 Township GREEN Formation i County ASHLAND D Clinton Se Well Ma 67892 Lease Name Classification ACTIVE WD ff Fig52 Result of adding a complete new well Warnings If API Number is not entered Gas Consortium d This Well Number is nat valid Enter again Warnings If API Number entered is already in the database 68 Gas Consortium x This Well Mumber is already in Database Note The dates for completion perforation stimulation amp well test should always be entered as the output of the software is directly dependent on the chronology of events The format of date is also specified for the user where required A close picture of that format is below Perforation D ate 1 1 05 1371 MODA Y You need to enter API well number only once in the well bore tab and it will be automatically co
56. overs re completions and re stimulation of the producing wells The parameters associated with these remedial operations can be controlled by the operator It is a fact that every well will respond to a specific remedial operation in a unique way For example the deliverability of well A will increase two folds if a proper restimulation 15 performed on it while the same operation performed on well B will result in little or no deliverability enhancement Same is true for workovers Finding the best candidate for restimulation or workover each year among the tens or hundreds of wells 15 a challenging task Consider another situation where well will have a 70 increase if a restimulation 15 performed but it would have 65 increase if a far less expensive workover 15 performed Obviously performing a workover instead of a restimulation on well would be more economical this year EXECUTIVE SUMMARY Portfolio management a common practice in the financial market 15 essentially an optimization problem that attempts to increase return on investment The objective of this project is to apply the state of the art in optimum portfolio management to the gas storage field in order to optimize the return on investment associated with well remedial operations Each year gas storage operators spend hundreds of thousands of dollars on workovers re completions and re stimulations of storage wells in order to battle the decline in
57. ow Chart All Well Tests WELL TEST ANALYSIS TOOL The user can do three types of Well Test Analysis in this software 1 Simplified Analysis for calculating n PD rate amp AOF 2 LIT Analysis for calculating Skin if k is known 3 Build up Test Analysis If Detailed Multi Point Test data is available 79 Lechnelogy Consortium FekiMameRBCD IN m tees paraa nayss Test Type khim 123 1 init 250000 0680 Tews Shin 1 216411 ADF MciD Fig64 Well Test Analysis button The analysis tools are very similar for Simplified and LIT Analysis except where mentioned The interface below will appear when you select Well Test Analysis button It will give you a glimpse of what has happened on the well since it was drilled WELL TEST ANALYSIS gt gt 12 345 67890 SmeliedAnsyss TIT Ames 5 5 5 58 JAP Number Dae restope kh Skin PeskDayRate AGF Detailed MP Data 12 345 57890 5 12 1965 Open Flow 12345 57890 11 76 1965 Opern Fim 12 345 57890 4 76 1966 Open Flow 12345 57890 4 24 1969 Open Flow 12 345 57890 3 25 1963 pen Flow 12 345 57890 7 12 2004 _Multi Point Double click the well test in the Grid that you want bo select for Simplified Analysis or LIT Analysis 2 Single click in the Det
58. pied in the rest of tabs and procedure is the same for editing The format for entering well API Number is API Well Mo 12 345 67880 If wrong format or well API number is entered for a new well then you will be greeted with the following message Gas Storage Technology Consortium 1 j Well Number Format is naE correct Correct Format 2 2 842 Adding only well bore completion perforation stimulation well test data l Click the Add New button Add ew while keeping that tab active for which you want to add the data Only for well bore tab following message pops up 69 Gas Consortium vou want to Add only Wellbore Data 7 Otherwise you must know the dates of Stimulation completion Perforation amp Well Test Click Yes e button to add only Well bore data The background color of all text boxes of that tab will be yellow indicating that they are ready for entering data 2 Enter the data The dates for completion perforation stimulation amp well test job should be known 3 Click the Save button Editing data One can edit complete well or just only a new well bore completion perforation stimulation well test data by following methods Editing a complete well data 1 Click on the Edit button while keeping your well bore tab as active 70 Following screen pops Gas Consortium B xj e vou want Edit only wellbore Data Otherwise vou must
59. port 15 available to public free of charge One important note about the software 15 that the current publicly available version of the software includes a neural network model that has been developed for our industry partner based on the data that they made available Once a storage operator decides to implement this software they should contact the principal investigator of this project Shahab D Mohaghegh Professor Petroleum amp Natural Gas Engineering West Virginia University Email shahab wvu edu Tel 304 293 7682 ext 3405 Web Site http shahab pe wvu edu and arrange for development of a neural network model for their specific storage field In order to make the best use of capabilities of the software package it 1s recommended that the storage filed have a minimum of 75 wells wells with data that can be used for analysis EXPERIMENTAL No experimental work was performed during this project 10 RESULTS amp DISCUSSIONS This is the detail report of the progress made so far in the above mentioned project which consists of following components 1 Project Overview 2 Data made available and its format 3 Neural Network Model 4 Genetic Optimization Model 5 Database amp Software PROJECT OVERVIEW The objective of this project is to apply state of the art intelligent optimum portfolio management to the gas storage field in order to optimize the return on investment associated with well remedial operations Columb
60. r Perforation Data Perforation data retrieved from a file Microfiche to Database process Different formats of Nitrogen Amount Data addition and refinement for Stimulation Data _ Tubing head pressure profile for multi point test Bottom hole pressure profile for multi point test Flow test 1 Delta pressure squared vs time Flow test 2 Delta pressure squared vs time Extended flow test Delta pressure squared vs time raph Gas properties simulator Calculation of true skin Calculation of true skin from build up test Retrieving flow rate of an open flow test Flow Diagram of Well Test Analysis procedure Neural Network Inputs and their source Accuracy of training data for the Neural Net Accuracy of calibration data for the Neural Accuracy of verification data for the Neural Net Different option in the software that make it versatile Data addition and refinement for well test data Screen shot of database showing different tables Main Screen of software File Mam options Vasa Screen shot o
61. rgest source of stimulation data came from the thousands of microfiche with some data being found in well summary reports Following 15 a picture of data in well summary reports 24 amp gung FornatorelRecon Log amp Date in Commerts D TE Bguvme J TTD Rs RS Water Fracture 7 1 1868 abr lyp dear Frackre a 2031 orsi b TT 11 Pumpe ied eraak dew 2500 8 3e press 244 3 EP 1000 4 10 500 bs jBuild Up Tests Enno 1 E ERES j TI ___ ___ O O I 4 hanse J p Pd 44444 pepe E 0 a e E mpera and frac Historical comparizen cf curve indicates somita kas of delivery Analysis of fw besi nion possible weny ameguogs FH in swell Coke tubirez in original and Trac wel gr rooazs LUCAS Report run date Tuesday February 22 2003 Page 2 ofa Slim Track Marne Est Pere Press psig ipei Avg Treat Press psi Aw BH rate bpm Instant Shut in Pressure psi Flush Fluid Density ppa Fesser
62. rom the file menu then the program will ask the user to select the filled Template file from the location The new data will be appended to the existing data Remove all data from database If the user doesn t want to append the data to the previous database but instead wants to up load a whole new data then there is an option in file menu as highlighted in the snapshot below This option will remove all data in the previous database After removing the data from previous database the user can up load the updated data from the template or enter it in the software r Gas Storage lechnology Consortium Fig36 Remove all data from database 54 Exit The program can be exited by two options One is to exit by using the file menu and selecting Exit while the other is to select the cross on the top right corner as in normal windows based applications Exit Options Gas Storage Techrology Consortium acg Help Create Templ te Import from Filled out Template Mp LATA PEDE OAL ANAS Canditate Selection West Virginia University Fig37 Exit form file menu Help Another option that can be accessed from the menu bar on top of the main menu screen 1s the Help menu option User Manual Formulas Canditate Selection NG West Virginia University 2006 Fig38 Help menu options 55 It contains three types o
63. s Estimation of n C PD rate amp AOF Same as above except that the points used are from the same test NOTE The n C PD rate amp AOF values for more than 400 well tests were manually calculated 31 Estimation kh skin true skin non darcy coefficient l From extended draw down test plot Pi Pwf vs time on log log paper Draw unit line for un stimulated wells and half slope line for Stimulated wells Find end of well bore storage effects after 1 1 2 log time cycle 2 Find values of viscosity z factor compressibility of storage gas at different pressure assuming Gas gravity 0 585 amp temperature 75 535 Draw Down Test l Plot Py vs time 2 Draw straight line after pseudo steady state starts 3 Find slope and P 1hr ji kh 1637qTzu m Pp k 5 SELIS Etu gl Jen m LCT 6 Plot skin vs flow rate It should be a straight line 7 Slope of this line is D 8 Find True Skin S at q 0 Build Up Test l Plot P vs tptdt dt on semi log paper 2 Draw straight line after well bore storage effects diminishes 3 Find slope and 1 4 163 7 8 9 5 PP iog 6 Plot skin vs flow rate It should be straight line 7 Slope of this line is D 8 Find True Skin S at q 0 We require time flow rate amp Bottom hole pressure from the data which are present in two txt files as bottom hole amp surface recording files The flow
64. s shown below inertial turbulent flow 5 Calculate b from equations below AW 2 2 4 q 2 AY SC YMa Ya NY AV Ya E 0 2 d Ze Plot bg 5 4 on a logarithmic scale The transient data points should form a N Number of data points b straight line If they don t form straight line calculate and b again with the data which forms the straight line Calculate Skin 5 with formula je a x10 S 3 23 0 869 1 632x10 T Where AY Delta Pseudo Pressure k Effective permeability to gas md h Net pay thickness ft t Flow time hrs Porosity Initial Viscosity Ci Initial compressibility psi 1 T Temperature of the reservoir R Tw Well bore radius ft S Skin dimensionless 41 Flow Diagram of Well Test Analysis procedure Following is the flow diagram of the well test analysis procedure and the type of values that we get from the data Well Test Analysis FLOW DIAGRAM t YES AN Fi 2 gt fit kis pven i Wer Box Do you want to continue 7 naa Msz Box you want to contine Find AOR POR Fig23 Flow Diagram of Well Test Analysis procedure 42 RESERVIOR CHARACTERISTIC It includes some reservoir properties The complete list of the data type retrieved is shown below API Number Well Radius Reservoir Porosity Reservo
65. should be drawn on the Horner plot on the left side of the end of well bore storage line shown in green on Horner plot 89 T WELL TEST ANALYSIS gt gt 12 345 6 7890 Input Data Permeability Analysis Diagnostic Plot T hrs Click the first point which does nat Form unit slope line Pi psi si 719 Flow Rate Redraw Diagnostic Plot Draw Slope by holding left mouse button down Pseudo pressure Press after slope is drawn Log Horner Time hrs Draw slope on the left side of green line Redraw Hi Horner Plot Plot Permeability md 1582 Skin 1 713 Claculate skin Save Fig75 Calculating skin nom Hornet plot After the slope is drawn the user can select Calculate Skin button to find the skin of the well The respective graphs can be redrawn any time by selecting the Redraw Diagnostic Plot or Redraw Horner Plot The value of permeability and Skin can be saved in the database by selecting Save button 90 SELECT WELL DATA In this form the user can choose to select the data of the wells that he wants to look at Following are a few ways he can choose the data Selecting a well by State County The user selects the state first and then the county All the wells will be selected for that county in the selected wells list box SELECT SELECTED WELLS OHIO COUNTY API Well Number W 34 005 01272 34 005 105
66. t box 1 Now the user can select the well tests that he she intended to the simplified analysis on them Then a line should draw based on the selected well tests in the picture box Figure 68 The slop n will be calculated by mouse clicking on the Calculate n Button The values of C Peak Day Rate and Absolute Open Flow will be calculated and shown in the grid as shown in the picture shot on next page These results can be saved in the database by selecting the Save button WELL TEST ANALYSIS gt gt 12 345 67890 Well Tests Simplified Analysis 2 1245 67990 5 12 1965 Open Flow wv 3 12 345 B7830 11 8 1855 Open Flow PHOT SRM IPC IPETTE versi 4 24 1353 pen Flow b 12345 67990 3 23 1853 Open Flow 12 345 B7830 7 12 2004 Multi Point lag Delta P sgr log Rate i Slope of line Draw Points gt gt Clear gt gt n 127220432 lt lt Calculate Selected Well Test PDrate AUF 2 12 345 67990 5 12 1855 Opern Flow 0 000222817039 13665 3 12 345 67890 11 8 1355 Open Flow 0 00021 3591 399 13467 Results Calculate PDR AOF Fig68 Simplified well test analysis tool 83 Note If there is only well test before or after the stimulation then the value of n can assumed and written in the textbox in front of label n as shown in the picture below The value of n cannot be assumed for more than one well
67. t with left mouse button down Fig73 Selecting the build up section from pressure profile If by mistake draw down data is selected then the following message will appear informing the user to select build up data again WELL TEST ANALYSIS gt gt 12 345 67890 X Wr Ne 1 Select Build up Test 88 After the portion of build up data has been correctly selected the permeability analysis tab will show following graphs The first one is the log log diagnostic plot between Del Pressure and Del Time The user should select the first point which does not fall on the unit slope line drawn by holding the left mouse button down The initial pressure Pi and flow rate text box values will be selected from the build up portion of the extended well test selected The graphs will be drawn again with a green line drawn on the Horner plot indicating The End of Well Bore Storage tewbs WELL TEST ANALYSIS gt gt 1 2 345 6 7890 en Input Data Permeability Analysis Diagnostic Plot Select the first point which does not fall on a unit slope line Del T hrs does nat Farm a unit slope line Pi psi 719 67 Flow Rate MefD 625 Redraw Diagnostic Plot 4 Pseudo pressure Slope should be drawn to the left of this line Log Horner Time hrs a Draw slope on the left side of green line Bip E Permeability md Skin Claculate Skin Fig74 Diagnostic plot analysis The slope
68. te permeability for the welltest where YES is written vi eli Tesl Type Fig71 Multi point well test analysis 86 If the cell with value YES 15 selected the software will read the data from the excel file and progress bar will become visible like in the picture shot above showing that the data 15 being read After the complete data has been read by the software the following screen will appear showing the pressure profile of the well test WELL TEST ANALYSIS 12 345 67890 Input Data tp low praductian time hrs 2 Temperature Weallbore Radius ft 1 Specific Gravity Porosity 72 0 1 4 Thickness ft m Le T a La LL Time hrs Select the Extended Buid up Test with left mouse button down Fig72 Well extended pressure profile All the Input data is retrieved from the database and if it is not found then default values are inserted The value tp flow production time is 2 hrs by default but can be changed by the user The Extended build up test for 2 or more hours should be selected by keeping the left mouse button down The green lines will indicate portion of build up test selected 87 F WELL TEST ANALYSIS gt gt 12 345 67890 E 1 Permeability Analysis tp flow production time hrs 2 Temperature F wellbore Radius ft 1 Specific Gravity Porosity 22 0 1 4 Thickness ft Pressure Time hrs Select the Extended Buid up Tes
69. the analysis and provide the list of wells and their corresponding remedial operation for each year based on the budget constraints identified by the user The state of the art 1n intelligent systems application that 1s currently being used extensively in the Wall Street 15 the methodology to achieve the objectives of this proposed project This methodology includes a hybrid form of artificial neural networks genetic algorithms and fuzzy logic Columbia Gas Transmission Corporation is the industry partner of this project and cooperated with the research and development team in order to ensure successful completion of the project The software application that 1s the deliverable of this project and is explained in much detail in this report 15 available to public free of charge One important note about the software 15 that the current publicly available version of the software includes a neural network model that has been developed for our industry partner based on the data that they made available Once a storage operator decides to implement this software they should contact the principal investigator of this project Shahab D Mohaghegh Professor Petroleum amp Natural Gas Engineering West Virginia University Email shahab wvu edu Tel 304 293 7682 ext 3405 Web Site http shahab pe wvu edu and arrange for development of a neural network model for their specific storage field In order to make the best use of capabilities of the softwar
70. ulti Port Close Fig89 Type of the wells test in optimization process 99 When all the parameters for GA have been selected and user selects the Apply GA 1 button then the screen below will appear showing the values of optimized stimulation slurry and change in skin due to this stimulation The picture below shows the GA optimization done on well 12 345 67890 T Candidate Selection Lat Long Water Acid Gel Foam He SumFluids Sandua Pror kh After TestT yp Cost Del Skin 1 20 62 A227 2 03 1 1 243 7 1855 28022 24848 35225 123 1 4 901 294678 1 921 Fig90 Optimization process for one well 100 If the optimization 15 applied to all the wells then we can rank the wells according to the change in skin by selecting Rank the wells from Options menu bar on the top left corner of the form as shown in snapshot below Candidate Selection Options To water Acid Gel Foam N2 SumFluids SandQuail Prio kh After TestT yr GA After Skirl Cost Del Skin 675 1134 121 85 25277 14157 469 67 123 5 024 458092 2 044 068 1052 161 08 7 16204 17397 5107 63 395 4959 595277 11013 338 617 15375 27294 164 96 6340 51 378 5048 575179 1 3871 405 11 21 10303 4509 1189 32 23835 62 100 4457 389857 1 7965 108 571 7845 7 20586 9666 328757 5233 4982 301995 1 2657 O G6 12447 7 301543 135 84 880 63 362 5048 470491 09724 1215 646 23481 137 55 255 06 117 42 210 504 855
71. ural net to interpret the well test With changing price of hydro carbons the petroleum industry is going through fluctuating material cost The stimulation material prices change frequently and are a factor of demand and supply in that region The software has the option to change the price of the stimulation material before applying the GA to the available data 48 Analysis of raw data vs refined data WELL TEST DATA DATA FIELD 8 INTIA L DATA S225 2223 3223 3224 3223 3222 347 347 347 52 292 292 222 AVAIL ABLE FIN AL DATA AVAIL ABL E Total wal I T ati AFI Hu m b r Ham d Pre oa 1 Flowing Pre aa FlaldFra 117 Flowing Pre 112 Ald Fra aad Flowing Pre 113 Pre s aura Est Rowing Pressure Est Absoluta Open Non Darcy Delta P Squared mem eene bo so B c h e 14 m as LEO ESE ad OLE ao ELE a cb SPT Ra ge APP S He JC a zy 2 j er er er E eg Number of Data Entries INITIAL DATA AVAILABLE DATA Fig29 Data addition and refinement for well test data 49 DATABASE amp SOFTWARE SOFTWARE BASICS This software allows you to add edit well data in the database and choose the data that you want to look at for a selected well It also has a Well Test Analysis tool which calculates the well deliverability parameters like n C Peak Day rate amp Absolute Open Flow T
72. wn from this research 1 The Artificial Intelligence Tool can predict Skin with high degree of confidence 2 The Portfolio Management for re stimulation candidate selection provides a cost effective method for taking full advantage of annual budget for remedial operations d This software 15 the first successful attempt to combine Data editing Well Test analysis and Artificial Intelligence in one software package 102 REFERENCES NONE 103
73. y the results from this software elect cnntrolahie parameters acea is Water injection Sbi F v Tem sm AT afin F Fuam IL Acid X hires pma F Gel r Bui Baca dm 4 Sene GA Input Titles Cost 2 Import Valor T Charactenstic Expat the Griel To Ecce GA Charaterictics gt i i 1 Sahat bpi TAAT ipee Population Size s00 UU TANE well Test 1 Crossover Rake Y 60 En s xu Mutation Faba ha C of in C MulPan NN Fig28 Different options in the software that makes it versatile Cime One of such variability introduced is that the software can use any other neural net if it 15 required The option menu of the optimization screen has the option to import any other neural network Plus there is an option to select the available controllable parameters for the GA For example if the user does not want to use or does not have foam and nitrogen then he can unselect them as shown in Figure 3 18 The length of GA will change according to the selection As the Neural Net has Well Test Type as its input so the Select Well Test Type menu option gives the user an option to choose the test the user wants the ne

An Intelligent Portfolio Management Approach. Final Report

Contents

Download Pdf Manuals

Related Search

Related Contents