When a borehole becomes a stress sensor
Published in Social Sciences, Earth & Environment, and Civil Engineering
A simpler way to estimate in-situ stresses underground
Measuring the natural stress field underground is one of the most difficult and expensive tasks in geotechnical engineering. Yet every tunnel, shaft or deep excavation depends on knowing it.
Traditionally, the reference method has been hydraulic fracturing. It is direct and reliable, but also slow, costly, and operationally complex.
Our recent work explores a simpler question:
What if the borehole itself already contains the answer?
Full article:
https://doi.org/10.1007/s44290-024-00036-4
Reading stress from deformation
When stresses concentrate around a drilled hole, the circular section rarely stays circular. The borehole subtly flattens, developing breakouts and an oval shape aligned with the stress field.
This deformation is not random.
Decades of rock mechanics show that:
• elongation develops perpendicular to the minimum horizontal stress
• breakout width reflects stress concentration
• geometry encodes both orientation and magnitude
Instead of forcing the rock to fracture, we can simply measure this geometry.
Using six-arm calipers and televiewer logs, we reconstructed the borehole cross-sections and calculated:
• stress orientation
• σH/σh ratios
• horizontal stress magnitudes
All from shape alone.
Does it really work?
The key test was comparison with hydrofracturing profiles.
Across several argillaceous formations in Spain, the results showed:
• very close agreement in stress orientation
• consistent estimates of stress ratios
• reasonable magnitudes when cohesion is well constrained
In some cases, predicted directions differed by only a few degrees from hydrofrac measurements.
This is important.
Because hydraulic fracturing remains the gold standard, but it is also:
• expensive
• time-consuming
• dependent on fracture-free intervals
• difficult at depth
Ovalization analysis, by contrast, uses logging data already collected in most boreholes.
Where it performs best
The method is not universal.
Our results indicate it works best when:
• boreholes are deep enough for clear breakout development
• rock mass cohesion is known from laboratory tests
• multiple sections are analysed, not single intervals
• high-quality caliper or televiewer tools are used
Under those conditions, the borehole effectively becomes a passive stress sensor.
No injection. No packers. No induced fractures.
Just geometry.
Why this matters for engineering
For tunnels, mines and underground works, stress estimation is often limited by budget and logistics.
A method that is:
• faster
• cheaper
• minimally invasive
• and still reliable
can change practice.
Ovalization analysis will not replace hydrofracturing everywhere. But it can significantly reduce the number of tests required and provide continuous stress information along the borehole.
In many projects, that trade-off is decisive.
A broader perspective
There is also a conceptual lesson.
Sometimes the most useful measurements are already embedded in the system.
Instead of adding complexity, we can learn to interpret what the ground is already telling us.
In this case, a small deviation from circularity becomes a map of the underground stress field.
And a borehole becomes an instrument.
Follow the Topic
-
Discover Civil Engineering
This is a fully open access, peer-reviewed journal that supports multidisciplinary research and policy developments across the field of civil engineering.
Related Collections
With Collections, you can get published faster and increase your visibility.
Waste-Based and Sustainable Concrete Materials: Performance and Structural Applications
The construction industry is under increasing pressure to reduce its environmental impact while maintaining the safety, durability, and performance of civil infrastructure. In this context, waste-based and sustainable concrete materials have emerged as promising alternatives to conventional concrete through the incorporation of recycled aggregates, industrial by-products, agricultural wastes, supplementary cementitious materials, and other environmentally friendly constituents.
This Topical Collection aims to provide a focused platform for recent advances in the development, characterization, and application of waste-based and sustainable concrete materials. Particular emphasis is placed on material design, mechanical properties, durability performance, microstructural behavior, structural response, and practical applications in sustainable construction.
The Collection welcomes original research articles and review papers addressing experimental investigations, numerical modeling, performance assessment, durability evaluation, life-cycle considerations, and structural applications of sustainable concrete materials. Topics may include, but are not limited to, recycled aggregate concrete, waste ceramic concrete, fiber-reinforced sustainable concrete, geopolymer and alkali-activated materials, supplementary cementitious materials, low-carbon binders, durability under aggressive environments, structural strengthening, and performance-based design of sustainable concrete elements.
By bringing together contributions from researchers worldwide, this Collection seeks to advance knowledge on environmentally responsible and high-performance concrete technologies and to support the development of resilient, durable, and sustainable civil infrastructure.
This Collection supports and amplifies research related to SDG 9, and SDG 11.
Keywords: Sustainable Concrete, Fiber-Reinforced Concrete, Waste Materials, Recycled Aggregates, Cementitious Composites, Durability, Mechanical Performance, Structural Applications, Sustainable Construction, Advanced Concrete Materials
Publishing Model: Open Access
Deadline: Apr 30, 2027
Machine Learning Applications in Smart Materials: Toward Sustainable and Resilient Infrastructure
The growing challenges of climate change, urbanization, and resource scarcity demand innovative approaches to designing infrastructure that is not only intelligent and adaptive but also environmentally sustainable and socially responsible. This Collection in Discover Civil Engineering aims to showcase the latest research at the intersection of machine learning (ML), smart materials, and infrastructure engineering to promote systems that are more resilient, efficient, and aligned with the United Nations Sustainable Development Goals (SDGs).
We invite original research articles, reviews, and case studies that demonstrate the transformative role of ML in the development, characterization, and application of smart materials for future-ready civil infrastructure. Key areas of interest include, but are not limited to:
· ML-assisted predictive modeling of smart material properties and behaviors
· Sustainable design and optimization of smart construction materials (e.g., self-sensing concrete, phase-change materials, piezoelectric composites)
· ML-driven sensor fusion and data analytics for structural health monitoring and predictive maintenance
· Data-driven lifecycle assessment of smart materials contributing to reduced carbon footprint and resource efficiency
· Integration of ML in adaptive infrastructure systems for climate resilience and disaster mitigation
· Hybrid frameworks combining physical models with ML to enhance performance and interpretability
· Applications of deep learning, reinforcement learning, and transfer learning in infrastructure material systems
· Case studies demonstrating measurable impacts on sustainability, resilience, and SDG alignment
This Collection seeks to build a platform for interdisciplinary collaboration among civil engineers, material scientists, environmental researchers, and data scientists. Contributions should aim to demonstrate how ML can accelerate the deployment of smart materials to meet global sustainability goals and build resilient infrastructure for the future.
This Collection supports and amplifies research related to: SDG 9, and SDG 11.
Keywords: Machine learning, smart cities, sustainable infrastructure, SDGs, urban resilience, digital twins, intelligent transportation, energy efficiency, climate adaptation, predictive maintenance
Publishing Model: Open Access
Deadline: Jan 31, 2027
