Daily production optimization at Petrobras Marlim

Daily production optimization at Petrobras Marlim

This page is a business case describing a case study or a pilot project involving one or several industry partners

Average: 4.5 (2 votes)

2 users have rated this content. We would love to have your vote as well. Log in and rate!

Since 2010 and up to today, IO4 has been working together with Petrobras to further develop and implement the IO Center SmartOpt concept to automatically find the best production strategy for their Marlim P-35 FPSO. The goal is to create a recommendation functionality that can be added to their current decision support system. We believe that such functionality potentially increase the production with 1-2% by helping the production engineers to utilize the bottlenecks in a best possible way.

Marlim is a subsea field located in the north-eastern part of the Campos Basin, approximately 110 km offshore Rio de Janeiro State, Brazil, with a varying water depth of 650 to 1050 meters. Production started in 1991, and today there are approximately 100 production wells and 50 injection wells. The field holds 8 operating FPSO units. It is expected that the Marlim field will be producing oil and gas until 2030.

The reservoirs are known to be water driven, thus water replaces the oil within the reservoir as the oil is extracted. As a result the WCs are gradually increasing and the GORs are changing. Furthermore, as the Marlim field is gradually becoming a mature field, the reservoir pressures and production rates are declining.

One of its operating units, P-35, is a floating production, storage and offloading platform (FPSO), which consist of one topside manifold, two subsea manifolds and in total 17 wells. Seven wells are connected to subsea manifold 1, three wells to subsea manifold 2, and seven satellite wells are connected through the topside manifold placed on the platform, as shown the figure. Two parallel pipelines, one production and one test, connects the subsea manifolds to the platform.

To determine the optimal production strategy in day to day operations of P-35 is difficult. The asset has several bottlenecks, and balancing the tradeoff between these, with the goal of squeeze through as much oil and gas as possible, is non-intuitive. Their in-house simulator Marlim-II is used for decision support, both to monitor and to perform what-if studies. During the what-if study, the production engineer is essentially conducting a “manual optimization”, testing alternative strategies and choosing the best one based on a predefined goal, in this case maximum production of oil. 

In this pilot project we utilize several modeling and optimization techniques, jointly called SmartOpt, to automatically online, in a few seconds, generate optimal recommendations. In short we can say that the method utilizes the production network structure by splitting the simulator into one component for each well, pipeline, compressor, separator, etc. Instead of connecting the components inside the simulator, we are essentially connecting them inside the optimization algorithm. By doing so, we find the optimal well routings, choke settings and gas-lift rates for all wells, so that total production of oil is maximized. The method takes into account that the pressure drop through the system should be feasible and that gas available for gas-lift is limited and needs to be distributed among the wells. Further, it also takes into account that the handling capacity of gas, water and liquid is limited and make sure that this is obeyed. 

The method finds the optimal solution to this problem in less than 10 seconds, which are two orders of magnitudes faster than the current state-of-the-art methods available on the marked. This enables a totally different production engineer workflow, since instant re-optimizations could be conducted on-the-fly during e.g. a well optimization meeting. It essentially automats the what-if study, and enables the production engineer to focus on other important questions like; what is the preferred operational envelope for each well, and is the inflow performance ratio predicting the well behavior accurately for the recommended (optimal) production strategy, etc.

When the pilot project is finalized, Petrobras should have a field proved method that should be ready for implementation within their own decision support system. We believe that this new functionality potentially could increase the P-35 daily production rate with 1-2% by helping the production engineers to utilize the bottlenecks in a best possible way.

Other key information

207 results
Below, you will find related content (content tagged with same topic(s) as this business case)
Content type: Ongoing activity

IO Aspects for Safe and Efficient Drilling

New methods and work processes for anomaly detection

Content type: Topic page

IO Capabilities and Capability Development

Combination of human skills, processes, change and technology, building capacity and ability to decide, plan and execute according to business

Content type: Video

IO Center

A short video introducing the IO Center

Content type: Ongoing activity

IO Center annual report for 2013 is now published

The annual report covering the activities in 2013 is now available.

Content type: Ongoing activity

IO Center participation at SPE Intelligent Energy

Several of the IO Center researchers and industry partner representatives presented papers at the SPE Intelligent Energy conference in Utrecht.

Content type: Report

IO MTO analysis - Optimal allocation of work tasks and competence within a distributed organisation

MTO analysis method gives practical guidance on how to perform analysis and verification of new organisational concepts.

Content type: Ongoing activity

IO13 material online

The presentations and other material from the IO Conference are now available online.

Content type: Publication

Joint Optimization of Oil Well Placement and Controls

In this work, we propose a joint approach that embeds well control optimization within the search for optimum well placement configurations.

Content type: Publication

Kinematical characteristics of the factorized model

This is a scientific publication written in collaboration with the IO Center.

Content type: Publication

Kinematical characteristics of the factorized velocity model

This is a scientific publication written in collaboration with the IO Center