All Earth

Deep Learning in Geomechanics and Geoengineering

Recent advances in deep learning have transformed data-driven research and engineering practice across a wide range of scientific and technological fields. In geomechanics and geoengineering, the increasing availability of experimental, monitoring, imaging, and numerical data provides new opportunities for applying deep learning to complex problems involving rocks, soils, underground structures, and geological environments. Deep learning techniques can complement conventional analytical, experimental, and numerical approaches by enabling efficient feature extraction, pattern recognition, prediction, parameter identification, and intelligent decision-making. Their applications are rapidly expanding in areas such as rock and soil characterization, mechanical behavior prediction, geotechnical monitoring, damage and failure identification, underground engineering, geological hazard assessment, and infrastructure safety. This Article Collection aims to bring together recent advances in deep-learning methods and their innovative applications in geomechanics and geoengineering, while also highlighting challenges related to data quality, model interpretability, generalization, and integration with physics-based knowledge.

Geomaterials such as rocks and soils are inherently heterogeneous, anisotropic, and often discontinuous, resulting in complex mechanical responses that are difficult to characterize accurately. For analytical tractability and computational efficiency, conventional theoretical and numerical approaches commonly rely on simplified assumptions, such as homogeneous or equivalent-continuum representations. Although these approaches have provided the foundation of modern geomechanics, such idealizations may not fully capture the spatial variability, nonlinear behavior, and complex failure mechanisms of natural geological materials, potentially leading to discrepancies between predictions and actual engineering responses. The rapid development of deep learning provides a promising data-driven pathway for addressing these challenges. By extracting complex features and nonlinear relationships from experimental, monitoring, imaging, and numerical data, deep learning can complement physics-based approaches and enhance characterization, prediction, and decision-making in geomechanics and geoengineering. Exploring their effective integration is therefore important for advancing more accurate, intelligent, and reliable geoengineering practices.

This Article Collection will focus on recent advances in deep learning and their applications in geomechanics and geoengineering, with particular emphasis on two interconnected directions:

1. The first concerns the integration of deep learning with conventional experimental and diagnostic techniques in rock mechanics, including image-based characterization, acoustic emission analysis, ultrasonic testing, digital image correlation, damage identification, mechanical parameter prediction, and intelligent interpretation of laboratory and field measurements.
2. The second focuses on deep-learning-assisted large-scale modeling and simulation in geoengineering, including surrogate modeling, reduced-order prediction, multi-scale analysis, uncertainty quantification, and accelerated numerical computation for complex underground and geotechnical systems. Contributions addressing physics-informed learning, multimodal data fusion, model interpretability, transfer learning, and generalization across geological conditions are also encouraged.

The Collection welcomes both original Research Articles presenting methodological or engineering advances and Review Articles synthesizing recent developments, challenges, and future research directions in deep-learning-enabled geomechanics and geoengineering.

Keywords:

1. Deep Learning
2. Geomechanics
3. Geoengineering
4. Rock Mechanics
5. Data-Driven Modeling

­­All manuscripts submitted to this Article Collection will undergo a full peer-review; the Guest Advisor for this Collection will not be handling the manuscripts (unless they are an Editorial Board member).

Please review the journal scope and author submission instructions prior to submitting a manuscript.

The deadline for submitting manuscripts is May 28, 2027.

Please contact Kara Roberts at kara.roberts@taylorandfrancis.com with any queries on discount codes regarding this Article Collection.

Please select Earth Observations from the list of available sections during submission. Failure to select the appropriate Article Collection or Section name can result in delays.

All Earth accepts the following types of article:

• Research Article
• Review Article
• Comment
• Rapid Communication
• Data Note
• Method
• Registered Report

Dr. Renbo Gao is currently a postdoctoral researcher at Southwest Jiaotong University and received his Ph.D. in Engineering from Chongqing University. His research interests focus on rock mechanics and underground engineering, with particular emphasis on interdisciplinary applications of deep learning. His current research investigates the stability and mechanical behavior of surrounding rock structures in underground energy storage systems, as well as deep-learning-based methods for rock mechanics testing and characterization.

Dr. Cheng Lyu is a Lecturer at Southwest Jiaotong University, China. His research focuses on rock mechanics and underground energy storage, with particular interests in the geomechanical behavior of rock masses and related engineering challenges. His work aims to support the safe and efficient development of underground energy storage technologies.

Dr. Zhuhong Wang has a Ph.D in Geotechnical Engineering at Chongqing University, China. His research focuses on AI-driven DIC/DVC, physics-informed neural networks, and intelligent computational mechanics. His work aims to advance artificial intelligence applications in geomechanics and geoengineering.

Guest Advisors declare no potential conflicts of interests, in line with our Editorial Policies.