Case Study (East Texas Gas Wells)
Now that we have actual producing assets, then what? Big talk about what could be done, but now time to start delivering on the Pontem vision. In this case study, we will focus on our East Texas (gas) assets. We were able to evaluate production data from 100+ wells to quickly assess potential optimizations, both pre-and post- acquisition. Initially, the challenge was simply collating production data from various sources (hard copies, PDFs, Excel spreadsheets) and visualizing everything in a concise way. At this stage, we aren’t even sure what we are looking for, beyond patterns in the data. This can all be done…but it requires data. And time. Patterns emerge with more data, but ideally we would like to use domain knowledge to close the gap on time to acquire data. With large data sets comes opportunities to “see” things, but also a potential “data blindness” that obscures what is the underlying causation of the results. Curve-fitting is not data science, or so I have been told. But initially - because we had a large data set - the focus is not on the “why”, but more on the “what”.
Once visualized, we began to leverage our domain knowledge and start to bridge the data gap to determine what variables are having a tangible impact on production. Again, from a pure data science perspective, this requires data. A lot of data. And (unfortunately), the ideal data set may have periods of time with failures, sub-optimal performance, and inefficiency. Used correctly, however, these transient or non-conformance periods actually provide some insights into what is occurring in the field.
Given enough time / resources, we know operators and service providers will get around to optimizing production and increasing value. Given all of the other demands, best intentions can go by the wayside and the biggest fire will get the attention. So, our question has always been - is there a better (quicker) way? We believe there is.
Our goal is to not only leverage data to understand what is happening, but combine with domain knowledge to explain why this is happening. To do this, we need to understand fluid chemistry, integrity management, well design, multiphase flow hydraulics, process engineering, and a host of other domain-specific parameters in an effort to put the observed data into context. All of which exist under Pontem Analytics and are available to the Pontem Minerals division.
In this example, we knew that there is a significant operating cost associated with chemicals. A necessary cost that is built into the operating expenses, but one that can potentially be optimized through a more focused lens. And the aim does not always have to be net cost reduction. The real benefit could be a reprioritization on which wells to treat first, allowing operating teams with limited resources to be more effective. So, we opted to have an eye on how to prioritize the wells. Within the asset team, there was a belief that if the wells were treated more frequently (and given some additional TLC), the production rate would increase. And, that was true…sort of…
This is the danger of blanket statements: Certain wells DID respond (~50%), but that also means ~50% either had no change or had a negative correlation to the chemical treatments. Understanding which wells responded to chemicals - and what made those wells unique (condensate yield, completion depth, tubing size, proximity to take-away capacity, etc.) - allowed a more integrated and efficient chemical management program to be implemented.
In other cases where the chemical program was not deemed to be materially relevant, other parameters dictated. In this example, we were able to quickly assess a key variable (shut-in duration / build-up time) and fed that into the overall reservoir management strategy.
As more domain knowledge of the wells was incorporated - often in conjunction with conversations with the operations team on what their experience had shown (but was never “written down” as a procedure) - a comprehensive well management plan was developed / implemented. Having focused conversations with everyone was key, letting the data help drive the conversation and connecting it to the human experience in the field, but not always dictating the answer. An empathetic use of data, if you will. And was successful in a majority of cases.
And even for the cases where it was not successful, we learned some things. In some cases, “doing nothing” was a great answer and it allowed the operators to focus attention in other areas. Sometimes, getting to NO is just as good as getting to YES. It gives similar insights.
From an overall asset perspective, the field rates were increased as a result of implementing these recommendations.
As more data is collected, the recommendations (or “algorithms”, to sound more authentic) are also updated, so that we can fine-tune the operations. Not only does this result in increased production and decreased operating expenses, there are other tangible benefits such as reduced manpower traveling to sites (improved safety / reduced carbon footprint).
Conclusions
At Pontem Minerals, we look to not only challenge the status quo and ask in-depth questions, but provide guidance on solutions to increase enterprise value of our assets. Our unique combination of domain knowledge in chemicals, production, and operations…combined with extensive data analytics capabilities…provides the bridge between collecting, analyzing, and monetizing data.
Be sure to check out our website for Pontem Minerals, which showcases more about our areas of expertise.