On December 2nd, 2021 Bardasz was invited to participate in a keynote conference in the 6th SPE Colombia Digital Oilfield Annual Meeting, a high-level technical event organized by the SPE Colombian Section whose aims to share with the Oil & Gas community those advances, work, innovation, and projects related to digital transformation and the use of digital tools for energy transition and operational efficiency, key challenges for our country.
In this event, our CTO Mark Farnan presented the conference "Challenges of Real-Time Analytics while Drilling", where to define concepts related to the systematic computational analysis of data or statistics and showed the most important seven challenges in real-time analytics. Here, we describe the following challenges:
#1 – What is “Real-time analytics” and how is it different from historical?
Analyze the drilling job whilst it is actually happening & gain timely insight to improve the outcome at the right time and use historical data to improve the confidence level of the prediction over time.
#2 – What to analyze in real-time and why?
What should we get the computer to analyze in real-time for us? The first one is detecting something bad has occurred, and alerting operators to it. The Last one is trying to predict that something bad might occur and warn us in advance. We could translate this in four points:
- Safety: limits breached and Failure event detection
- Efficiency: Rigstate (or Opstate) Determining rig activity from the data Rig KPIs. (Connection times, plan vs actual)
- ROP Optimization
- Event Prediction: stuck pipe prediction, equipment maintenance, and others.
#3 – Getting reliable streaming drilling data
Here, Mark explained the importance of data transfer monitoring, the energy industry data standards, monitoring for data stoppages, dealing with poor rig-site communications, gap handling, and loading missing data. What data do I need? Where do I get it? Data Frequency and Latency. Understanding requirements for frequency and latency of the data is critical to a successful real-time analytics project.
#4 – Streaming Realtime data problems
Not all our data is time-series. A significant amount is indexed by depth, and the collection time is ‘lost’ at the source. At this point, it defines the data governance, quality assessment, mnemonic standardization, curve mapping, sensor ‘noise’ / outlier detection, and filtering.
#5 – Building a streaming algorithm
What language are you going to write it in?
What Development environment?
Can you ‘run’ the algorithm with streaming data in the dev environment?
#6 – Running real-time analytics in production
Challenges when running oil-field analytics on existing “big data” stream processing systems.
#7 – Actionable intelligence – Alerts & Alarms
The purpose of real-time analytics while drilling is to be able to take action in a timely manner, during the operation. This point described the actionable intelligence, what happens with nuisance alarms, and highlighted the importance of delivering the right information to the right person, at the right time.