In our previous blog post we provided a high level view of why oil & gas industry leaders chose Tamr to be their data unification solution. We hope this post on Wells Unification (or Wells Harmonization within the industry) can help oil & gas experts get into the details about our capabilities and we hope data experts not familiar with this industry can also benefit from the post.
Why do we need a unified view of oil wells?
Oil wells are expensive to develop (even after you are lucky enough to know where to drill). According to the U.S. Energy Information Administration (EIA), onshore wells typically cost between $4.9 million and $8.3 million, including costs related to land acquisition, capitalized drilling, completion, facilities costs, lease operating expenses, and gathering, processing and transport costs.
Oil wells are expensive to operate. Producing oil and gas involves completing the wells, connecting them to pipelines and then keeping the flow of the hydrocarbons at an optimal rate, all integrally related to the subsurface environment — the type of rock and structure of the reservoir.
Thus, for each well we invest in, we want to get the most out of it and do our best to operate it. Analytical approaches that impact the success rate of finding or reducing the cost to develop and produce oil and gas can make energy more affordable, safer and environmentally conscious.
Today, every well that’s drilled uses extensive machinery, measurement devices and people – all of which produce huge amount of structured and unstructured data — in terms of the volume, variety and velocity of data being captured.
Why is wells harmonization so hard to achieve?
First of all, the variety of data types and data sources. Wells data records are generated from multiple internal sources throughout the well life cycle. There are data from the planning stage, drilling and completion stage, production stage, etc. For example, the drilling and completion data will have subsurface data that need to be used in the production stage, and the data might reside in completely different systems.
Secondly, the data quality issues. Many oil companies have been operating for decades. Their legacy data might be recorded in different systems with different accuracies. A lot of the time, the data is simply missing. One is example is that the wells dataset might not have an API number (a unique identification number to represent each wells in different granularity) — a typical data quality issue seen with many Tamr projects.
Thirdly, extremely costly upkeep. To make things worse, companies usually jointly invest in wells development, and oil wells are bought and sold all the time (as well as oil operators themselves). It’s very hard to maintain good data governance when you are not even sure about ownership. Some of our customers need to spend a great deal of time and manual effort just to figure out their percentage share in a given well.
The Tamr wells harmonization solution
As a data-agnostic platform, Tamr Unify picks up signals from the entirety of the data and integrate them. Using human guided machine learning, the algorithms were able to learn from oil & gas experts without them having to do any coding.
We use information such as well names, location information (longitude/latitude), operation information (rock formation), etc. to match well headers, well bores and well completions that belong together with very high accuracy and speed.
Data Mastering: Gain 360-view of wells by simply answering yes and no questions. As an example with publicly available oil well data, Tamr Unify successfully mastered well header data and permits from a southern state government website to achieve an overall accuracy of 92% after only ~2.5 hours of model training.
Tamr Unify can also assist in identifying gaps and breaks in well header information across multiple applications. Once wells have been mastered across sources, Unify can also prioritize unmatched wells for further investigation.
Golden Records: Tamr Unify also helps generate a ‘Golden Record’ of wells data as a result of the harmonized sources, which can be leveraged to increase analytic accuracy, correct data-break errors at source and reduce null counts in well data improving well production modeling accuracy.
Tamr Steward: When a data issue is identified, the Tamr Steward extension can capture user feedback for follow-up by an admin. Tamr Steward captures information about the relevant data and metadata, and allows the users to describe the issue and tag other users. Open issues are documented and centralized for DataOps teams to take action.
Some key analytical outcomes
As we discussed earlier, high costs in oil well development and production mean that the large amount of science, machinery and manpower required to produce a barrel of oil must be done profitably, taking into account cost, quantity and market availability. The path to optimizing production and development is highly dependent on models created and data acquired through the well life cycles. Some key analytical outcomes include:
Timely decision making: On site teams can make operational decisions (such as what fracking method, what pumping fluid to use, etc.) based on the latest data with a thorough picture — currently using forecasted subsurface data or data from regulatory sources.
Better analytics with higher data quality: The company can easily compare individual well productivity against benchmarks across their portfolio and analyze investment return based on actual productivity; downstream data consumers can see the data sources and types in a more transparent manner,so that they can trust the data.
Reduced manual effort in terms of both preparing information from different sources and stages for data consumers (sometimes in Excel) and integrating a new data system, such as a better GIS data system.
Tamr envisions a world where people in large enterprises consume accurate, up-to-date, unified data distilled from many silos to deliver transformational outcomes. We believe data science and data engineering will help the oil and gas industry learn more about each subsystem and inject more accuracy and confidence in every decision, impacting the success rate of finding or reducing the cost to develop and produce oil and gas, and making energy more affordable, safer and environmentally conscious.