Innovative and Integrated Digital Solution for Water-Handling Management System Using Machine Learning for Production Enhancement in a Brownfield

Abstract Water management has become a challenge and a key factor to achieve proper field management in a brownfield in the Amazon region. Apart from historically increasing the basic sediments and water (BSW) and flow assurance events, the longevity, ullage, and lack of maintenance of the processing facilities have come key conditions to sustain production and reduce lifting costs. An integrated digital solution was implemented to enable predictive analysis and behavior of the water system at the facilities level, such as horizontal pump system (HPS) failures, line plugging, injection rate deviations, as well as forecasting of injection rates in real time to improve the efficiency of operations and avoid production deferment. Several failures had occurred in the water-handling system caused by the lack of real-time monitoring or fast event detection and corrective actions. This has led to many shut-ins and failures of electric submersible pumps (ESPs), and injectivity losses in the waterflooding and disposal wells, triggering production losses. Hardware for data collection in selected points and customized digital advanced workflows using data analytics and machine learning (ML) were implemented and developed to optimize processes and production. Therefore, using the connectivity provided by a satellite system, supervisory control, and data acquisition (SCADA) optical fiber and operations monitoring platform, the variables are now monitored in real time to enable early identification of events, give a rapid response, and optimize production of the field. The digital solution was implemented using the following steps: Data capturing: Data gathering was improved by adding sensors at key points, aiming to minimize data gaps in the field, and ensuring comprehensive availability in the database. Surveillance workflows: Multiple workflows were created using edge computing and ML to have an early alert and recommendations to make a rapid response. Opportunity identification: Workflows were implemented in real time, enabling early identification of events in the field to give a rapid response, debottlenecking, and optimize production by proper water-handling management through data shown in engineering dashboards. This innovative and integrated digital solution has shown outstanding results. Monitoring the data from the water-handling system in real time and applying engineering workflows has led to a 76% reduction in processing time, 75% reduction in commuting for data gathering, and 5 t of carbon dioxide (CO2) reduction per year. Due to the early event identification, the prediction of potential failures, and a rapid response time, the operational team has reduced deferred production by 283,000 bbl of oil, saved USD 9 million related to HPS failure reduction and saved production, which contributed to extending the ESP's run life (RF), optimizing maintenance costs, and reducing lifting costs. In this paper we outline how the implementation of a novel digital solution can optimize the process, production, and cost of the critical points for a brownfield by using data analytics, edge computing, and ML, potentiating the digital transformation path in all stages of the energy industry.