What do retail and mature oil/gas fields have in common? The answer: variate and complex datasets, tons of it. And here is when Data Analytics and its powerful methods and techniques come in handy!
The Pipeline:
Design and implement a customizable and SCALABLE-ELASTIC analytics SOLUTION; prepare and refine SILOED and or publicly available not fully exploited datasets from mature Oil and Gas Fields, and perform Visual Exploration, Descriptive Analytics, Predictive Analytics, and FORECASTING of oil/water production, etc.
The SLIDES show a few examples of possible designs: from the basic, using only free-of-cost Google Colab and Workspace collaborative tools, to a very sophisticated design pattern using several Google Cloud Platform services and tools.
Scalable-elastic means that cloud-based solutions like this automatically and dynamically adjust up and down smoothly to the workload and user demands, as illustrated in the image below.
Solutions like this are relevant for the Oil and Gas Industry because the actionable knowledge extracted from SILOED and (or) publicly available, not fully exploited datasets is an opportunity for any field operator to generate REVENUE in the short term with REDUCED investment costs.
The proposed solution can be used directly as a SCOUTING TOOL for asset acquisitions to quickly evaluate Oil/Gas blocks' potential placed up for bid or be further customized and used in other practical applications; results can be integrated directly into any Oil and Gas traditional workflow.
To use as a reference this example of implementation in the Google Cloud Platform (GCP), to learn how official YPF publicly available not fully exploited MASSIVE datasets (from January 2006 to April 2021) that comprise well data (name, latitude, longitude, total depth, etc.); and oil, gas, water productions, etc.; of more than 6 thousand wells from mature oil and gas fields in Neuquina and Cuyana Basins (Mendoza Providence, Argentina); data was cleaned, structured, refined and blended in Trifacta Cloud Dataprep; reading the data to carry out Advanced Visual Exploration, and Descriptive and Predictive Analytics.
The refined data is saved automatically in Cloud Storage, and directly accessed from Google Looker Studio to perform Quantitative Descriptive Analytics. In this example, the refined data was also published as a Google Sheet and connected with Tableau Public for Advanced Visual Exploration and Predictive Analytics: FORECASTING of Oil/Water production by AREA/RESERVOIR, etc.
A fully automated scalable-elastic version of this End-to-End Analytics SOLUTION can be implemented using, Trifacta Cloud Dataprep, BigQuery, Cloud Function, and Tableau Online. The figure above illustrates at a high level the solution architecture. It's important to say that the architecture mentioned in the previous two paragraphs and this fully automated version are both scalable and elastic: they can be fed by massive datasets and by small ones. As mentioned before, they smoothly adapt to the workload and user demands. Data Analytics is not restricted to so-called "Big Data" at all!
Results:
Implemented Design Pattern End-to-End Analytics Solution.
Easy-to-digest, Fully-Interactive Visualizations, and DASHBOARDS served in Tableau Public; actionable insight can be quickly extracted in just a few clicks.
To improve the User Experience, results are also served in Google Looker Studio as Easy-to-Digest, Fully-Interactive Charts, and visualizations. The actionable knowledge can be quickly extracted in just a few clicks; filtered data tables can be downloaded or published as Google Sheets and used in other applications or integrated into several oil and gas workflows.
The figure below depicts one of the dashboards implemented in Google Looker Studio. All solution-compelling visualizations were designed and implemented to meet the client's requirements and enhance the user experience so that the analyst could extract the critical insight needed with only a few clicks.
Similar criteria were used to build the visualizations in Tableau Public Software. The four images below show a set of fully interactive visualizations designed and successfully deployed. In this platform, it is possible to construct visualizations and dashboards with more sophisticated and advanced features, including forecasting using oil and gas production by month. Also, important information (see the image to the right in the two figures immediately below), such as polygons corresponding to surface facilities, 3D seismic surveys, seismic lines, water, oil, gas pipelines, etc., can be displayed and analyzed.
As seen from the last two figures above, where filters have been applied to zoom in and highlight a particular reservoir or field of interest (Ugarteche), it is evident that it has increased its oil production (curve thickness) while reducing water production (curve color less blue). The dotted trend line (a polynomial fitting) and the forecast consistently predict that the previously mentioned fluid production will be sustained in the next couple of years or so. This insight is critical if this field is placed up for bid.
After the analyst using the previous resources and procedure, identified an attractive reservoir/field (Ugarteche, for example), which other reservoirs or fields would be of interest? With this in mind and to accelerate the whole process of decision-making, an easy-to-use Content-Based Recommendation Engine was also built (using all the relevant data available) and included in the solution. The image below shows the engine's front panel implemented in Google Looker Studio. A Pearson Similarity Index is used to propel it. Reservoirs/fields from both basins (Neuquina and Cuyana) are displayed in the word cloud.
About the figure below, typing Ugarteche-Cuyana in the dropdown labeled Reference Reservoir(s), then setting the Similarity Index slide in 91% you got the top-5 reservoirs/fields (compelling displayed in the word cloud and ranked from top to bottom in the table to the left of the figure) more similars to Ugarteche reservoir. If Ugarteche (in Cuyana Basin) is attractive for a bid, the reservoirs La Ventana and Rio Viejas of the same basin would also be of interest. So the analyst, with just a few clicks, has available a short list of additional reservoirs of interest on which to focus, saving time and labor even if more data and further analysis are required, accelerating the decision-making process.
Hoping the presented example of Analytics Solution, which includes several Google Cloud Platform Services and Tools, helps to streamline processes in the Scouting Tools realm. Want to learn more or carry out a One-on-One user test?
In a future post, I’ll consider a use case where the issue to address, the operator seeking to optimize the budget of a recently acquired mature Oil field, is to predict, out of thousands of wells, which are more likely to be successful if opened to production. Stay tuned, and don’t miss it out. Leave your comments below and share.
Commenti