Production Allocation of Oil and Gas: A case Study of the Skarv Field Simen Sæten Petroleum Geoscience and Engineering Submission date: June 2015 Supervisor: Curtis Hays Whitson, IPT Norwegian University of Science and Technology Department of Petroleum Engineering and Applied Geophysics Abstract Oil and gas production will typically involve wellstreams from numerous sources being brought and processed at a common facility. Allocation of hydrocarbons has an impor- tant part in the petroleum industry, and it can be defined as the procedure of assigning the portions of the commingled stream to the contributing wells. This is done by using all the measurement equipment available at the field, which might include pressure-gauges, multiphase meters, test separators, and fiscal meters. The operating conditions at which the measurements were obtained will be different for every instrument, depending on the location of that instrument. This thesis will investigate how the allocation procedure is currently done at Skarv (offshore field operated by BP), and propose an alternative allo- cation procedure. Implementing a proper solution for the allocation system can be difficult, as it requires a vast amount of information to be connected and manipulated. One of the most important things to consider is how the measurements are brought from the local line-conditions to standard conditions, as this is a vital part in determining what the surface rates actually are. At Skarv, the individual well rates are periodically measured with subsea multiphase meters, and these rates are converted to standard conditions by using simplified black-oil models on a volumetric basis. One of the challenges found by using this current method is that the calculated total oil rates out of the facility are systematically underestimated, compared to the measured oil rates into storage. The proposed alternative allocation is an automatic procedure created in Pipe-It and is based on compositional streams. The main advantage by using compositional streams instead of the simplified black-oil models, it that the compositions can be sent through any defined topside process. This allows us to do sensitivities of the process train and iii Production Allocation of Oil and Gas: A Case Study of the Skarv Field investigate its influence on surface rates. It was found in this thesis that using a detailed surface process, which includes scrubbers and recycling of the gas, would increase the calculated condensate rate by over 70% compared with using black-oil models. Doing the allocation using compositions will also allow us to investigate the effect of commingling. This effect was also found to be significant, and could result in a difference up to 6% in the calculated total oil out of the facility compared to processing the streams individually. iv Sammendrag Allokeringavhydrokarbonerkanbliansettsomenmetodefor˚afordelestrømningsratenfra et prosesseringsanlegg til de induviduelle brønnene. Dette gjøres ved˚a bruke de forskjel- lige m˚aleintrumentene som er tilgjengelig p˚a feltet, inkludert trykk-m˚alere, flerfasem˚alere, test separatorer og fiskale metere. Trykk- og temperaturbetingelsene vil være forskjellige for alle de ulike instrumentene, avhengig av hvor instrumentet er plassert. Det er derfor nødvendig ˚a konvertere ratem˚alingene fra de lokale betingelsene til standardbetingelser, hvor de s˚a kan bli sammenlignet med hverandre. Denne oppgaven vil undersøke hvordan allokeringsprosedyren er gjort ved Skarv (offshore felt operert av BP), og videre foresl˚a en alternativ allokeringsprosedyre. DeindividuellebrønnratenevedSkarvblirperiodiskm˚altvedhjelpavsubseaflerfasem˚alere, og disse ratene vil s˚a bli konvertert til standardbetingelser ved˚a bruke black-oil modeller. En av utfordringene ved˚a bruke denne metoden er at de beregnede totale oljeratene blir systematisk underestimert i forhold til de oljeratene som faktisk blir m˚alt ut fra anlegget. Den alternative allokeringenprosedyren er laget i et program kalt Pipe-It og er basert p˚a komposisjonsstrømmer. Den største fordele ved˚a bruke komposisjoner i stedet for de forenklede black-oil modellene er at komposisjonene kan bli sendt gjennom en hvilken som helst definert prosess. Dette gir oss muligheten til˚a undersøke hvilken innflytelse denne prosessen har p˚a de kalkulerte oljeratene ved standardbetingelser. En av prosessene som var undersøkt i denne studien inkluderete alle de ulike separasjonsutstyrene, i tillegg til resirkulering av gass. Ved˚a bruke denne detaljerte prosessen økte de beregnede konden- satratene med over 70% i forhold til verdiene fra black-oil modellene. Et annet resultat fra denne studien er at effekten av ”commingling” har ogs˚a en betydelig inflytelse p˚a de kalkulerte ratene. Denne effekten kan resultere i en forskjell p˚a over 8% i de beregnede v Production Allocation of Oil and Gas: A Case Study of the Skarv Field oljeratene ut fra prosessanlegget i forhold til˚a prosessere strømmene individuelt. vi Acknowledgments I would like to express gratitude and appreciation to my supervisor Professor Curtis H. Whitsonforhisguidancethroughoutthiswork. Hisexperienceandengagementhavebeen essential. I am also very grateful for all the help from Arnaud Hoffmann, who have acted asaco-supervisorandhelpedmeimplementingthealternativeallocationsysteminPipe-It. This project was done in cooperation with the BP team, and could not have been com- pleted without their help. I would especially like to thank Tim Griffin and Carole Ross for their assistance and support. Kareem Basha has also been of tremendous help and has showed me how the allocation system at BP is currently operated. I am of course grateful for everyone that have been of help, including Hugh Rees, Siagian David, and Ove Helge Soma. I also wish to acknowledge all the support from friends and family. Thank you. Trondheim, June 2015 Simen Sæten vii viii Contents Page 1 Introduction 1 1.1 Purpose of Allocation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 1.2 Objectives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4 2 Theoretical Framework 5 2.1 Fundamental Aspects of Allocation . . . . . . . . . . . . . . . . . . . . . . . 5 2.1.1 Typical Facilities and Interfaces . . . . . . . . . . . . . . . . . . . . . 5 2.1.2 Overall Material Balance for a Production System . . . . . . . . . . 7 2.1.3 Topside Process and Calculation of Surface Rates . . . . . . . . . . . 8 2.1.4 Comparison of the Different Measurement Systems . . . . . . . . . . 11 2.1.5 The Allocation Process and System Imbalance . . . . . . . . . . . . 13 2.1.6 Fields Using Periodic Well-Tests . . . . . . . . . . . . . . . . . . . . 17 2.2 Phase Behavior . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20 2.2.1 Modified Black-Oil Formulation . . . . . . . . . . . . . . . . . . . . . 20 2.2.2 Cubic Equations-of-State . . . . . . . . . . . . . . . . . . . . . . . . 23 2.2.3 Two-Phase Flash Calculation . . . . . . . . . . . . . . . . . . . . . . 27 2.3 Well Performance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29 2.3.1 Inflow Performance Relationship . . . . . . . . . . . . . . . . . . . . 31 2.3.2 Vertical Flow Performance. . . . . . . . . . . . . . . . . . . . . . . . 32 2.3.3 Natural Flow . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33 2.3.4 Estimating Rates Using Pressure Measurements . . . . . . . . . . . 34 3 Allocation at the Skarv Field 37 3.1 Production System at Skarv . . . . . . . . . . . . . . . . . . . . . . . . . . . 37 3.2 Processing Unit at the FPSO . . . . . . . . . . . . . . . . . . . . . . . . . . 39 3.3 Current Allocation Methodology . . . . . . . . . . . . . . . . . . . . . . . . 41 ix Production Allocation of Oil and Gas: A Case Study of the Skarv Field 3.3.1 Black-Oil Properties Used in the Allocation Process . . . . . . . . . 42 3.3.2 Results From the Current Allocation Process . . . . . . . . . . . . . 45 4 Alternative Allocation Method Using Compositional Streams 53 4.1 Pipe-It Software . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53 4.2 Reproducing Surface Rates Predicted by the Reservoir Simulator . . . . . . 56 4.3 Well-Test-Conversion to Compositional Molar Wellstreams . . . . . . . . . . 58 4.3.1 Description of Method . . . . . . . . . . . . . . . . . . . . . . . . . . 58 4.3.2 Seed-Feed Dependency . . . . . . . . . . . . . . . . . . . . . . . . . . 59 4.4 Influence of the Topside Process on the Calculated Surface Rates . . . . . . 66 4.5 Generating Lookup Tables . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73 4.6 Processing of Commingled Flow at the FPSO . . . . . . . . . . . . . . . . . 79 4.7 Recommendation and Future Work . . . . . . . . . . . . . . . . . . . . . . . 85 5 Conclusions 87 Definitions and Nomenclature 89 Bibliography 94 Appendix A Worked Example Using Uncertainty-Based Allocation 97 Appendix B Derivation of Rachford-Rice Equation 99 Appendix C Inflow Performance Relationship 100 Appendix D The Beggs and Brill Method 103 Appendix E Choke Models 105 Appendix F Component Properties for the Topside Process 107 Appendix G Calculated Daily Rates Using the Current Allocation 108 Appendix H Comparison of BO-values from PVTp and those from PipeIt116 Appendix I Generating Lookup Tables Using AWK Script 118 x