October 14, 2021
In this talk we give an example of successful application of pSeven product of Datadvance to a particular problem of optimization of the oil production stimulation technology. Hydraulic fracturing has been used routinely in the field since 1947 to stimulate oil (and, later, shale gas) production. The design of a pumping schedule is usually obtained with the help of commercial simulators based on first principles continuum mechanics, where the model needs validation itself. We took an alternative route based on real field data on reservoir properties, well, fracturing design and production. Digital database now includes more than 5000 wells from 23 oilfields of Western Siberia (Russia), with 6600+ fracturing operations in total. Starting with ~400 parameters characterizing each well, we end up with about 40 key governing parameters used as input features for forward modeling. An inverse problem (selecting an optimum set of fracturing design parameters to maximize production) is formulated as optimizing a high dimensional black box approximation function constrained by boundaries and solved with surrogate-based optimization from pSeven, which demonstrated better performance than alternatives. A recommendation system containing all of the above methods is designed to advise a production stimulation engineer on an optimized fracturing design.
We will also discuss the outlook for the future, where we need to optimize technologies with respect to the set of ESG metrics: minimize carbon footprint (Scope 1 emissions), environmental impact (reduce the amount of chemicals) and minimize solid material injected underground (Scope 3 emissions of supply chain). Net zero is a data integration and data analytics problem, where access to good quality data and advanced analysis instruments becomes crucial to provide a well-grounded advice to the industry.