Real-Time Data-to-Information Systems for Improved Decison- Making in Production Optimization

Transcription

1 Force Seminar April 21-22, 2004, NPD, Stavanger, Norway Real-Time Data-to-Information Systems for Improved Decison- Making in Production Optimization Jan-Erik Nordtvedt Managing Director Epsis AS

2 Buzz Word Compliant... Digital Oil-Field Smart Field e-field Real-time Reservoir Management Oil-Field of the Future i-field Smart Assets Integrated Production Management

3 About Epsis Epsis is a technology company targeting the realtime reservoir management market niche, focusing on data-to-information and services 2003 was Epsis first year in operation Epsis has currently 11 engineers engaged Epsis client list includes ChevronTexaco, Norsk Hydro, Shell, NFR, Statoil and ABB Focus: Evaluation of real-time reservoir management opportunities for a specific asset Real-time reservoir management technology development

4 Real-time Reservoir Management Brown fields: East Texas, onshore gas California, onshore heavy oil 3 NCS, offshore oil Argentina, onshore oil Green fields: NCS, subsea, DW gas GoM, subsea, DW oil

5 Real-time Reservoir Management Better Decisions Faster Decisions... and the ability to act on those decisions Technology work processes - people

6 The rtrm Vision and its Implications To optimize the flow of hydrocarbons from the reservoir to the off-take using a process control approach A paradigm shift not due to a single technology change Integration of existing and new technologies A key is converting data into information to obtain faster and better decisions This is demanding new work processes and skill sets Transforms the way we operate

7 10 Years From Now We Will: 5 NOW Operate in integrated and remote operation centers Continuously optimize production and injection along the well Have new prediction tools allowing us to act proactively Have completely different work processes and skill sets

8 Some challenges along the road Process Facility Completion Production Reservoir Monitoring Data management and IT infrastructure Analysis: Data-to-Information Optimization Work process automation / changes Change management

9 Decision Loops Operator Optimization Production Optimization Field Optimization Operator Optimization: Safety Control room decisions Minimize downtime Production Optimization: Choke & sleeve settings Identify well-work and lost production Well and reservoir surveillance Field Optimization: Optimize reservoir drainage, injection Schedule interventions Short loop: Process optimization Export obligations Second to hours Medium loop: Production optimization Accelerating production Hours to weeks Long loop: Integrating reservoir engineering and geology / geophysics Accelerating production and improving reserves Weeks to months

10 Data Swamp NNM gas condensate development 6 wells Data measured for each of the wells: RT Sand detection RT p,t downhole and top-side, and in firststage separator RT flow rates top-side RT choke-size monitoring 800 TAGs * 60 s/min * 60 min/h * 24 h/d = approx. 70 mill. datapoints datapoints / day

11 Data Utilization 60-80% of an engineer s time is spent on finding and preparing data We under utilize available data The technologies are ahead of our ability to utilize them

12 Data-to-Information System Technologies Data cleansing Work process automation for faster and better decisions: Dynamic well-test analysis Data mining / Neural networks / AI Visualization Faster and better decisions Faster and better decisions

13 Data-to-Information Systems Need focus on decisions and decision support information rather than raw data Need to utilize data measured routinely to update the well status so that possible problems and opportunities are identified early enough to remediate problems and capitalize on opportunities Avoid time consuming data manipulation to arrive at results Avoid excessive use of Excel... (escape from Excel hell ) Automate work processes were engineering judgment is not critical for the work process output

14 Data Cleansing Why? Time resolution of data is very different Analysis models require data with different characteristics Well-test analysis; high resolution after transient Decline curve analysis; weekly / monthly values What? Data reduction Noise filtering Outlier removal Transient detection

15 Wavelets Automatic noise filtering Noise removed from pressure data y-axis y-axis x-axis x-axis

16 Dynamic Well-Test Analysis We measure P wf downhole and production rates top-side as function of time Build-up test: The well is shut in and the pressure in the well is building up as a function of time We utilize a model of the flow in the reservoir to calculate formation permeability and nearwell formation damage (skin) from the transient pressure data We utilize the production history to determine the drainage area and reservoir pressure Output: k, S, P r, A Q P wf

17 Automatic well-test analysis Trigger: A well is shut in (any reason) TWP: A well-test module is run Pressure and rate data filtered, transient detected Model identified k, S, A determined RT PI information enabled D2I distribution: Information distributed to right personnel (if needed) Decision / actions: No action (everything OK), well treatment (problem identified) Trigger & Criticality Identification TWP D2I distribution Decision & Action SEQUENTIAL & MANUAL

18 Some Challenges Data-to-Information and Work Process Automation: Shift in focus: From analysis tools to using the output from the tools (i.e., more focus on decisions) How can we shorten the data-to-decision loop maybe by moving away from a sequential mind set? How can we ensure that we re not only getting faster decisions, but also better decisions? Reduce the time is spent on finding and preparing data (from 60-80% to 0?) Utilize relevant data, throw away irrelevant data Get into a proactive mind-set

19 Summary The North Sea has generally very good level of instrumentation and data infrastructure Process Facility Completion Production Reservoir New technologies are emerging within the production and reservoir domains for real-time data-to-information conversion Technology work processes people: There is a need to focus on data-toinformation, work processes (incl. automation) and change management to realize the value TWP The challenge is successful asset implementations (with appropriate metrics for measuring success)