Deviation Tracker

In the context of well operations, "deviations" refer to abnormal conditions or anomalies that deviate from the expected operational behavior. These deviations could range from changes in flow estimates, sudden temperature spikes, or unexpected well shut-ins. Detecting and addressing these deviations promptly is critical to maintaining optimal production efficiency and preventing costly issues such as equipment failure or production loss.

Problem Definition

Monitoring well operations for deviations is a complex task, often relying on engineers manually reviewing large amounts of sensor data. This process is not only time-consuming but also prone to human error, where subtle trends might be overlooked. Traditional monitoring approaches can generate numerous false positives, leading to unnecessary interventions or, conversely, missed detection of significant events like water coning or gas coning. Furthermore, communication between onshore teams and operators about these deviations often lacks structure, relying on informal methods like emails and phone calls, which can result in unclear status updates and delayed responses.

The Deviation Tracker Application addresses these challenges by providing a comprehensive, automated solution for detecting and managing deviations in well operations. It reduces the burden on engineers by automating the monitoring process, minimizes false positives through advanced data filtering, and ensures clear communication and documentation of detected events.

Solution

The solution consists of two main parts: a complex backend event generation and an intuitive frontend interface to enable inspection of the events for the users. The event generation is enabled with our own data mining framework (Squashy), where all intervals where the controllable variables are unstable can be filtered out. In the cases where control variables are changed, you expect changes in the system that are not deviations. However, when all control variables are stable and there are significant changes in the system, something unexpected could have happened. These events are exactly what the backend system catches and generates events.

There are three main groups of tag deviations; trend, instability and frozen signal. In addition more complex detection algorithms can be tailored for the asset, depending on what is relevant. One example could be detecting risk of water coning.

The dashboard displays all detected events for each well in a "Deviation Ranking" list. The ranking of the wells are based on a set of severity tags that monitors the degree of trending in the wells' production. For each event, associated information like type, start and end time is displayed. In addition, the user can mark events as relevant and add comments. When an event row is clicked, the event with the tag is plotted in the "Investigate" panel.

The Deviation Tracker Application exemplifies our commitment to delivering scalable, tailored solutions that enhance operational efficiency and decision-making in complex production environments.