Performance and Health Monitoring, Prognostics and Contingency for Electrical Completion Systems, Designed on Purpose
- Mihitha Nutakki (Schlumberger) | Marian Faur (Schlumberger) | Daniel Viassolo (Schlumberger) | Indrajeet Gour (Schlumberger)
- Document ID
- Offshore Technology Conference
- Offshore Technology Conference, 6-9 May, Houston, Texas
- Publication Date
- Document Type
- Conference Paper
- 2019. Offshore Technology Conference
- 6.2.2 Strategic Health Management, 4.3.4 Scale, 4.2 Pipelines, Flowlines and Risers, 7.6.6 Artificial Intelligence, 2.1.3 Completion Equipment
- Completions, Cloud, Prognostics and Health Management, Data Monitoring, Machine Learning
- 0 in the last 30 days
- 230 since 2007
- Show more detail
- View rights & permissions
Electrical intelligent completion systems are designed for reservoir management and production control of different zones within a well. The operating conditions expose the tools (hereinafter referred to as "stations") within the system to harsh conditions—high temperatures, high pressures, gas, and sand particles. Over time, exposure to such conditions can lead to component, module, station, or system failure, resulting in the possibility of deferred production or nonproductive time. To improve system reliability, the downhole and surface components of the system are equipped with sensors. The capability of such systems to acquire and transmit data related to the health of the system, its modules, and its components is a novel approach.
Acquired data is transmitted through a cloud-based Internet of Things (IoT) framework for station health monitoring and health degradation predictions. The high frequency of data acquisition in combination with a large number of stations can lead to huge volumes of data. Manual monitoring and processing of large-scale data can become very inefficient and unmanageable and consequently, there is a need to develop intelligent algorithms for processing data to make actionable decisions and to adhere to a sustainable workflow.
This paper describes the data pipeline established through cloud-based architecture for automating the monitoring, dashboard creation, and health prediction for the electric motor actuator (EMA) module, using historical health data of the downhole electronics. In addition, a predictive approach consisting of feature engineering, event (actuation) extraction, and supervised machine learning algorithms is discussed and illustrated through example data sets and results.
|File Size||920 KB||Number of Pages||8|