Implementing an Integrated Production Surveillance and Optimization System
You have access to this full article to experience the outstanding content available to SPE members and JPT subscribers.
The complete paper describes some of the technical challenges faced in deepwater operations and the methodology adopted for implementing an integrated production surveillance and optimization (IPSO) system to mitigate the risks. The IPSO (deepwater-digital-oilfield) system is designed to be a combination of emerging information and communications technologies, intelligent algorithms, and fit-for-purpose asset-management work flows.
The IPSO System
The vision of the digital oil field is to foster an environment where work is performed with fit-for-purpose tools and streamlined work flows to measure, model, and control the field by enabling people to make the right decisions at the right time. This is enabled through a seamless decision hierarchy (Fig. 1) that achieves integrated asset management across business processes, the hydrocarbon value chain, and temporal scales. This multiscale optimization is handled by dividing the challenge into subproblems and passing down decisions from each level as targets for lower-level processes, while system outputs are considered as feedback loops to correct actions.
A fit-for-purpose digital-oilfield system, IPSO, was developed to optimize base production through real-time surveillance and efficient reservoir management. The system automates routine tasks such as data capture, data manipulation, preliminary analysis, visualization, and data storage. IPSO was primarily designed as a decision-support system, with emphasis on value-driven work flows for producing assets. Experts from several fields are required in order to pull together a successful solution.
The business justification of the IPSO system is dependent on the production phase of the field, where different optimization opportunities exist. Across these diverse assets, there is an inherent risk of customizing the architecture, approach, and tools to such an extent that scalability and extensibility are threatened. Such customized solutions slow down the adoption of good features from one project to others. Therefore, a conscious attempt was made from inception to develop IPSO work flows with a common denominator that could be configured for different business needs across a company’s portfolio.
The IPSO system was deployed on line more than 2.5 years ago for a single asset and was later extended to all deepwater assets. Over this period of time, a number of multidisciplinary work flows were added on the basis of the foundational principles detailed in the complete paper.
The following six key tenets were used as guiding principles in developing IPSO work flows:
- Business value
- Effective planning
- Stakeholder support
- Involvement of a champion
IPSO Work Flows
Data Integration and Visualization. The common data tier is the foundation for all IPSO work flows. Data are seamlessly federated from the real-time historian, production databases, and other sources for automatic calculations and visualization purposes. By not replicating any source data, a single version of truth and reproducibility is ensured. Smart multiscale sampling algorithms are used to enhance speed and computational performance. A logical data model is used to provide a familiar environment for all end users across many assets.
Surveillance. The basic premise behind surveillance work flows is to support asset personnel making key real-time decisions with relevant contextual information. This involves the following:
- Aggregating data from underlying data sources (both real-time and historical)
- Filtering data as appropriate to balance accuracy with performance
- Validating data to ensure quality
- Detecting key events (such as shut-ins, well restarts, and choke changes) and abnormalities
- Performing routine calculations for analysis
- Visualization of results in customizable user interfaces
Analysis. It is imperative to identify and rectify any productivity anomalies as early as possible to reduce downtime and improve recovery. Recognizing in a timely manner that productivity has declined is often important and can be challenging when the responsibility is shared between multiple teams working with different metrics.
Pressure-transient analysis of the shut-in period (pressure buildup) is limited to a very small duration of the entire operation. Deconvolution techniques increase the analysis duration under certain conditions but are still anchored to a buildup period. Both of these methods require human interpretation and are not conducive to automation.
Recognizing that every flow period in a well’s life cycle has embedded information led to systematically assembling and analyzing a variety of methods used to compute well productivity. Every method has its assumptions, strengths, and limitations; therefore, looking at the results as a whole often helped in early identification of abnormal conditions and led to expedited actions.
A typical well’s life cycle includes stable flow periods, shut-ins, restarts and ramp-up sequences, and operating-point changes; the liquid rate and bottomhole pressure evolve during these periods. The following phases are analyzed within the IPSO system.
Shut-In Period. Every shut in (planned or unplanned) is automatically detected with multiscale, unsupervised algorithms. Once the start and end of the shut-in period are identified, buildup pressures at the end of 10 and 60 minutes are extracted as a proxy for true average reservoir pressure to compute a well-productivity index.
Ramp-Up/Restart Period. Following each shut-in, the well is restarted and ramped up to desired flow rates. While slow ramp rates may result in undesired production deferment, fast ramp rates could potentially cause formation damage. An automated algorithm was developed to, first, detect the start and end of the ramp period without any human intervention. As a post-processing step, maximum ramp rate and total ramp rate were calculated for each ramp period. Next, the ramp-rate calculation was computed in real time and presented to the operators as a dashboard while the wells were being ramped, along with target rates provided as operational guidelines from the engineers.
Operating-Point Changes. During routine operations, changes in choke settings (wellhead or facility inlet) and separator conditions are often made that lead to changes from one stable operating condition to another. This provides two points on a stabilized inflow-performance-relationship curve for the well.
Dynamic Flow Period. Deconvolution can be applied to recast this varying-rate period as a constant-rate period with its associated pressure response. However, in order for deconvolution to be valid, a number of conditions need to be reached, and validation of such analysis typically needs human expertise. Therefore, a machine-learning-based method is adopted to learn the reservoir model from real-time pressure and flow-rate data. A linear-regression model is proposed to predict bottomhole pressure from flow rate and time. By use of the “learned” reservoir model, a virtual shut-in experiment is then performed to predict bottomhole pressure for a constant-production-rate period followed by a 60-minute buildup. Shut-in analysis is then used to calculate the well-productivity index.
Reservoir Performance. With reservoir pressure calculated for each well from the shut-in analysis described previously, continuous reservoir-voidage calculations and pressure-monitoring diagnostics are generated automatically for each reservoir unit.
Optimization. In some mature assets, the declining reservoir pressure leads to competition between wells to flow through shared subsea infrastructure. Because there is no excess pressure available to choke back wells in such cases, this often leads to production loss if the flow rates are not balanced across all flowlines. Continuous well-routing optimization is then required with the most-recent well-potential data and the current available facility infrastructure. An integrated well and network model configuration based on current routing information is automatically constructed, and optimal routing along with target rates for each well is computed.
The following case studies illustrate examples in which proactive surveillance and reservoir management using IPSO have added tangible business value.
Smart-Sleeve Malfunction. In a deepwater asset, the producers were completed with sliding sleeves to control zonal production in a stacked reservoir with individual downhole-pressure and -temperature gauges for each zone. By detecting the departure between upstream and downstream pressures across the sleeve by use of the IPSO system, the engineer was able to identify a smart-sleeve malfunction quickly. The issue was resolved within 1 week, avoiding a costly workover to replace or repair the smart sleeves and downtime caused by smart-sleeve closure.
Reservoir Connectivity. During a greenfield startup, there was significant uncertainty in reservoir connectivity because of complex fault architecture. The flowback tests were inconclusive, and a reservoir-management plan was put in place to start up the wells in a strategic manner over a period of time to discern reservoir connectivity. The IPSO system was used to analyze the Well B shut-in, while pressure interference from the offset well (Well A) was picked up automatically. This information substantially changed future development and saved the cost of at least one infill well (approximately $120 million).
Productivity Decline. In another deepwater well, the shut-in-analysis algorithm identified productivity decline by automatically detecting shut-ins, calculating productivity indices, and tracking them over time. This allowed a low-cost pump-in treatment to be performed to remediate the ongoing productivity decline, thus avoiding costly rig-based acid stimulation and associated downtime.
Improved Uptime. One of the fields was started up with the IPSO system for routine surveillance and reservoir management and achieved approximately 97% uptime during the first 6 months of production. Among other factors, having an automated system to track performance and highlight issues before they caused downtime was very beneficial.
Metering Check. Wells with subsea multiphase flowmeters are automatically tracked against topside multiphase flowmeters. Any deviations beyond a preset threshold raise an alarm for the subsea engineer to look into the calibration of subsea meters and thus avoid rate-allocation errors over extensive periods of time.
Production Deferment. The machine-learning-based well-productivity-index-calculation method captures well-performance trends quite well compared with actual shut-ins. This reduces the need for planned shut-ins to analyze buildups and helps to avoid associated downtime. Such continuous surveillance saves time for the engineer and changes the surveillance method to management by exception.
Implementing an Integrated Production Surveillance and Optimization System
01 March 2018
Improved Monitoring System for Heavy-Oil Steam-Assisted-Gravity-Drainage Wells
The complete paper provides an overview of the development of fiber-optic sensing for steam-assisted-gravity-drainage (SAGD) applications, including a review of more than 10 years of work in development and field applications in western Canada.
Integrated Surveillance Offshore Turkmenistan
The complete paper describes a work flow in which wells and production networks in the Cheleken Block offshore Turkmenistan are automatically modeled daily with steady-state and transient tools and ultimately analyzed by the Cheleken Block Central Data Gathering System.
There are two ongoing trends in the industry. The first is an increasing wave of digitization. Second, piggybacking on the increased access to sensor and equipment data, analytics and artificial intelligence are on an exponential rise in the industry.
06 March 2018
06 March 2018
09 March 2018
07 March 2018