Innovative thinkers

Vol. 232 No. 9

Dr. Ali Dogru
The reservoir snapshot worth a billion barrels

Dr. Ali Dogru

Three decades ago, the inner workings of complex petroleum reservoirs were a deep mystery—one that billions of dollars and the dream of the next big discovery rested upon. Since that time, Dr. Ali Dogru, Saudi Aramco chief technologist, has helped unlock the mystery of the petroleum underworld and brought dozens of declining fields back to life.

After receiving his PhD in petroleum engineering and applied mathematics from the University of Texas, in 1981 Dogru worked at Core Labs Inc. and in 1981 set out to develop one of the industry’s earliest chemical flood simulators for a far east client.
Dogru and his team analyzed 2D aerial and cross-sections of the brownfield to build the flood simulator, which the client then used to develop an enhanced oil recovery (EOR) plan for the field. For the next two decades, low oil prices meant that chemical flooding was not used. As a result, when Asian companies began
a series of chemical floods in the early 2000s, the technology was still similar to the simulators developed in the 1980s.  

Early on, Dogru began to see the exponentially increasing gap between the volumes of reservoir data and the technology available to process it. Until the 1990s, single-computer CPUs powered Saudi Aramco industry-standard simulators, which required that large reservoirs be separated into large grid blocks. “Fields were huge … and data had to be upscaled significantly to fit into the simulator model. Upscaling led to the loss of most of the reservoir heterogeneity characteristics,” Dogru said. “By using more vertical layers we could capture vertical heterogeneity better, but results could not be totally representative since the flow in the third dimension was ignored.”

While using the vertical cross-section models to study water movement within the reservoir, Dogru had to introduce layers a few feet thick to better capture the water arrival times. ”I noticed that the more layers we added, the better we could simulate shale layers, and water arrival times got better and better,” Dogru said. Also, matching the water arrival and cut was not accurate, as the simulator model ignored tiny shale layers. “All this information led to my idea that we need to build finer-scale reservoir models, and we should develop a new simulator to handle millions of grid blocks,” Dogru said. “This could only be accomplished by parallel simulation.”

Based on the principle of some of his early parallel simulators, in 1998 Dogru led a team that developed a new system from scratch, named the Parallel Oil, Water and Gas Enhanced Reservoir Simulator (POWERS). In the early development phases, Dogru faced resistance from colleagues: “‘Simulation technology is here; why not buy a faster computer? Why are you re-inventing the wheel?’ they asked. I knew the wheel was not there ... I wanted to invent new wheels instead of reinventing the existing ones.”

POWERS was first tested on the giant Berri Hadriya reservoir using 1.3 million cells and 74 vertical layers. “It ran 30 years of history in four hours,” Dogru said. In 2002, Dogru was able to fit the entire Ghawar reservoir into a single model using 10 million cells.

For large fields, however, the aerial grid blocks were still too big to account for seismic data, which was available for every 25–50 m of the reservoir. “To simulate seismic-scale models, we would need a billion-cell—or giga—simulator,” Dogru said. The team launched this effort in 2005, and just three years later, the first giga-cell models for Saudi Arabia’s giant Ghawar and Safaniya fields tested successfully. GigaPOWERS officially launched in 2010.

To Dogru, the possibilities of GigaPowers are limitless. With the next simulation frontier—incorporation of graphical processing units (GPUs), traditionally used in video games—currently in testing, he stands by his lifelong engineering mantra: “Do not reinvent the wheel. There are always better wheels!”  

nell.lukosavich@worldoil.com