From Market Basket Analysis to Well-head Monitoring: Using Events to Increase Oil Recovery

Abstract

BP has a successful, limited scope deployment of a non-parametric capability called Event Detection and Association (EDA). The technology resembles ‘Market Basket Analysis’ tools used to carry out associative assessments of item sets. Variations from the original analytical approach have been found necessary before the EDA techniques could be used to solve real-world oil and gas problems. These changes are based on fundamental physical laws such as conservation of mass and the recognition that systems of relevance all contain time delays between the ‘items’ that are to be tested for association. EDA comprises part of a suite of tools collectively referred to as Top Down Waterflood Diagnostic and Optimisation (TDWF). TDWF is a top down reservoir simulation toolkit capable of developing a statistically valid reservoir representation based purely on operational data together with a limited set of configuration data (the x, y location of wells is required). EDA contributes an independent validation of the Capacitance Resistivity Model (CRM) using the same data further processed to extract causal events on the injectors and delayed-response events on the production wells. By studying the time based correlation of these input and output events the basic TDWF CRM model is either confirmed or has suggestions made for improving its predictive capability. Early studies on three reservoirs, when combined with insights from suppliers of four different data-driven technologies, confirmed the robustness of EDA. A limited trial deployment followed in four operating regions. The users’ studies have generated insights and improved the product design and capabilities. EDA can deliver insights into well connectivity in the subsurface. Working from data alone EDA can even approximate the findings of a much more expensive tracer study and deliver these insights in a very short time and at a fraction of the expense.