2024 GMiS Luminaries

Education: Masters in Finance, CEMA University (Argentina), Executive MBA, IAE Business School (Argentina); Masters in Data Analytics, Texas A&M University; Masters in Petroleum Engineering, University of Southern California; Masters in Gas Business Management, Simón Bolívar University (Venezuela)

Over the course of his nearly fifteen years at Chevron, Alejandro Antonio Lerza Durant’s experience ranges from reservoir engineer to management advisor.  He’s applied novel technologies normally found in computer engineering and biotechnology, to increase his productivity and impact, including machine learning, multivariate data analysis and Artificial Intelligence. 

Alejandro’s application of machine learning to production optimization for  Chevron has reduced the time and boosted the accuracy with which petroleum engineers can estimate the value of a well.  Machine learning can inform both the selection of well locations and the optimum way to complete the wells – two of the most relevant factors the petroleum industry uses to calculate the profitability of wells. 

On one project, Alejandro applied complex data analytics and machine learning techniques to understand and predict the well production performance in unconventional reservoir formations.  He was able to identify key geological parameters of these kinds of reservoirs that were previously thought to have only minor impacts. As a result, he recommended optimal well locations and completion designs to increase the economic return of each well by 10 percent.  This translated to a project present value increase of $40 million. Alejandro’s work was later improved, applied to new field developments, and published for others in the industry to continue building on.  This particular publication confirmed how data analytics and machine learning are even more suitable for optimal reservoir development and planning than previously understood. This is just the latest in a body of 12 peer-reviewed research works, which has seen some 125 citations in the industry, a significant impact to be sure. 

By applying these novel technologies to raw data gathered by geologists and other scientists at extraction sites, he was able to not only generate a three-dimensional area model, but also better predict which other factors most impact the ability of a well to produce over time. The next step of the research is to build an even more powerful model that generates well production predictions from minimal data. In this context, Alejandro has proven to be among a very few pioneers at the intersection of machine learning and petroleum engineering.