Pre-drilling Lost Circulation Prediction in Special Lithology-developed Intervals of the Southern Dagang Exploration Area
This paper addresses the challenge of lost circulation in the Southern Dagang exploration area's complex volcanic and bioclastic limestone formations by developing and validating an integrated pre-drilling prediction framework that combines well log reconstruction, seismic inversion, and ant-tracking fracture characterization with a decision tree-based risk classification system to accurately identify high-risk intervals and optimize drilling operations.
Original paper licensed under CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer
Drilling for oil and gas is a high-stakes endeavor that takes machines deep into the Earth's crust, where the rock is often unpredictable. One of the most persistent and costly problems engineers face is "lost circulation," a situation where the heavy liquid pumped down the well to control pressure and cool the drill bit simply vanishes into the ground. Instead of circulating back to the surface, the fluid disappears into hidden cracks or porous pockets within the rock. This can halt operations for days, inflate costs significantly, and in the worst cases, force the abandonment of an entire well. While this issue occurs in many geological settings, it becomes particularly dangerous in areas where the rock itself is unusual, such as regions containing ancient volcanic flows or fossil-rich limestone. In these specific environments, the rock is not just a solid block; it is a complex, sponge-like structure riddled with holes and fractures that can swallow drilling fluid instantly.
In the southern Dagang exploration area of China, a team of researchers from PetroChina Dagang Oilfield Company set out to solve this puzzle before a single drill bit ever touched the ground. They focused on a geologically complex region where volcanic rocks and bioclastic limestones—rocks made largely of fossilized shells and coral fragments—are scattered among more common sedimentary layers. These special rocks are notorious for their hidden dangers. The researchers analyzed data from 204 previously drilled wells to understand exactly why fluid loss happened so frequently there. They discovered that the problem was not random. Instead, catastrophic fluid loss occurred when two specific conditions met: the presence of these special, porous rocks acting as a storage tank, and a network of tectonic fractures acting as high-speed highways connecting those tanks to the surface. Without the fractures, the porous rock would hold the fluid; without the porous rock, the fractures would not hold enough volume to cause a disaster. It was the combination of the two that created the risk.
To predict where these dangerous combinations existed, the team developed a new way of looking at the Earth using seismic data. Seismic surveys work like a medical ultrasound for the planet, sending sound waves underground and listening for the echoes to map what lies beneath. However, in this complex area, standard methods failed to distinguish the dangerous rocks from the safe ones because their acoustic signatures were too similar. The researchers solved this by creating a more detailed picture. They first used a technique to reconstruct missing data from the well logs, effectively filling in the gaps to create a clearer profile of the rock types. They then combined this with a multi-parameter seismic inversion, a process that uses advanced computing to translate the raw seismic echoes into a 3D model of the rock's physical properties. This allowed them to see the volcanic rocks and bioclastic limestones with much greater clarity than before, identifying their exact location and thickness.
Next, the team needed to find the cracks. They used an algorithm called "ant tracking," which mimics the way a colony of ants finds the most efficient path to food. In this digital context, the "ants" move through the seismic data, following the subtle discontinuities that indicate fractures. This method was sensitive enough to spot small-scale cracks that standard imaging missed. By fusing the map of the special rocks with the map of the fractures, the researchers created a single, comprehensive view of the subsurface. They could now see exactly where a wellbore would pass through a thick layer of porous rock that was also riddled with fractures. Based on this combined data, they built a decision-making system that acts like a traffic light for drilling. It assigns a risk level to every section of a planned well path. If the system detects a thick layer of special rock with intense fracturing, it flags that section as high risk, predicting a high probability of lost circulation. If the rock is safe or the fractures are sparse, the risk is marked as low.
The team tested this new approach on 16 newly drilled wells. The results showed that the method could accurately predict the risk levels before drilling began. In most cases, the system correctly identified where lost circulation would occur and where it would not. For the new wells drilled in 2025, the prediction accuracy reached 83.4%, a significant improvement over previous methods that often failed in this specific geological setting. The researchers noted that the system is not perfect; it cannot see fractures that are smaller than the resolution of the seismic waves, and it does not yet account for every engineering variable, such as the specific properties of the drilling fluid. However, by focusing on the geological reality of the rock and the fractures, they have provided a robust tool for engineers to design safer, more efficient drilling operations. This work offers a clear path forward for managing one of the most expensive hazards in oil exploration, turning a blind guess into a calculated prediction.
Drowning in papers in your field?
Get daily digests of the most novel papers matching your research keywords — with technical summaries, in your language.