Abstract

Structural geological models are used to understand and visualize how geological materials are organized below the Earth surface. This technique is crucial in different fields of geomechanics such as mining to identify mineral deposits and to plan mining operations, environmental engineering to assess the contamination of soil and groundwater and to design remediation strategies, and civil engineering to design foundations, tunnels and other underground structures. In this sense, ground models allow the engineer to understand and to visualize the spatial organization of subsurface geological structures and, additionally, it is possible to evaluate the spatial distribution of material properties which is essential in any geomechanical study prior to any project. In this context, SAALG Geomechanics through DAARWIN web platform provides a suitable software tool for the geotechnical engineer to perform the complete ground model workflow. The process starts introducing borehole data from different sources and then, the engineer can visualize and interpret borehole logs, define geomechanical units and assign parameters from in-situ and laboratory tests, as well as assign units’ geometry from geological layers. Finally, the user can feed this information to the system to obtain a 2D cross-section, that can be exported to the commercial Finite Element software PLAXIS and use this PLAXIS model for geotechnical design and more advanced analyses such as Sensitivity analysis and Back-Analysis, which are also implemented in DAARWIN.

Abstract

Building a ground model through manual processes can be time consuming, as large amounts of data need to be classified to define the extent and spatial distribution of the different soil materials. This paper delves into the application of machine learning (ML) methodologies, in conjunction with in-situ geotechnical testing data, to develop the ground model for a downstream dam founded on both weak and liquefiable soils. The dam covers a linear extent of approximately 800 m and was extensively characterized by means of in-situ tests, including 206 cone penetration tests (CPTu), 37 boreholes and 35 test pits. The performance of two unsupervised ML clustering algorithms are compared: Density-Based Spatial Clustering of Applications with Noise (DBSCAN) and an extended version with a hierarchical component (HDBSCAN). The clustering uses CPTu data, which consists of the normalized cone tip resistance (Qtn) and the normalized sleeve friction (Fr) varying with elevation. Nearby borehole logs are used to evaluate the results of both clustering methods for a single single CPTu sounding using different clustering parameters. Then, a global clustering including several CPTu soundings is done and results are compared with the ground model that was manually made using Leapfrog software. Both methods show very good performance, with HDBSCAN being better and more robust.

Abstract

There is an increased need for submarine cable infrastructure across the world to serve the offshore wind power industry and the expansion of the submarine interconnectors and telecommunication networks. Cable landfall projects are complex as land, intertidal, nearshore, and offshore environments all coverage, creating an array of dynamic processes, constraints, hazards, and engineering challenges that an asset may face over its lifespan. Robust site characterisation and ground modelling is key for the success of these projects, to allow for the effective, safe, and economical site selection, design, installation, and operation of an asset. In order to accomplish this, the integration of engineering and geoscientific datasets, obtained through several data sources and survey techniques is required; as well as collaboration and integration between multiple technical disciplines. This paper lists the key stages, techniques, and sources available for obtaining the required data. We then consider how the data can be managed and integrated to obtain a holistic ground model for use in the design, construction, and operation of the asset. We discuss the value these models can provide throughout the lifecycle of landfall projects.

Abstract

The further development of offshore windfarm areas in various countries plays a key role in the transition of energy production towards renewable sources. As offshore windfarm areas tend to expand and the amount of ground truth data is limited, the estimation of geotechnical parameters at unknown locations integrating other site investigation data becomes a necessary tool. This is especially relevant for cost efficient area wide site characterization. Here, the proper integration and correlation of geotechnical and geophysical data is a key factor for reliable ground model building. This study investigates different prediction methods, while presenting a modelling framework which incorporates geological, geotechnical, and geophysical information to derive synthetic Cone Penetration Testing (CPT) profiles using offshore windfarm site investigation data from the German North Sea. We combine geological interpretation, CPT data and 2D ultra high-resolution seismic reflection data. The geophysical and geological information are used to guide geotechnical parameter prediction. Additionally, seismic horizons constrain the prediction as structural information. For evaluation, we test and compare several prediction techniques, with different level of complexity, from geostatistical methods to machine learning. Seismic attributes are used as auxiliary information to improve CPT parameter prediction. To validate the results, CPT parameters are predicted onto a representative 2D seismic line and a leave-one-out cross-validation (blindtest) is performed. Though all methods struggle to replicate local extremes, results indicate a reduction of prediction uncertainty when implementing seismic attributes.

Abstract

Sometimes, when working in TSF safety analysis, the historic archives with the original ground topography or details about the TSF design and construction are few or inexistent. The knowledge of the bedrock position, as the details about the embankment construction are essential to know, for instance, the tailings thickness and the construction type. For the analysis of TSF current state stability the ground/structure model is essential, and is, in many cases, very difficult to define with few and low-quality data. The original ground topography is, often, obtained from aerophotogrammetric reconstruction, from satellite images or aerial photos, originally with low resolution and uncertainty of more than 10 m for the elevation. In this paper is presented a case-study from a TSF where the initial data package had only the feasibility design and some very simplified reports that checked the stability of the dam before an upstream raising, i.e. it didn’t exist much information and the details about the site and about the structure were very limited. To “add” difficulties to the process of defining the bedrock surface and to establish the TSF design, the embankments were built with local rocks (mainly schist and phyllite), and during the initial analysis of historical satellite images it was noticed that the original ground was excavated in different areas to increase storage area and obtain construction materials. This paper presents the steps developed to establish the definition of the bedrock ground surface and the difficulties felt and its impacts on TSF safety assessment are discussed.

Abstract

The construction time and cost of a rock tunnel project are highly dependent on the rock mass quality and encountered ground behaviour. In most rock tunnel projects, the knowledge about the ground conditions along the tunnel is limited, making it difficult to predict accurately the construction time and cost. The KTH model takes a probabilistic approach to address this problem; however, it does not account for the spatial variability of the ground conditions. This paper investigates an alternative probabilistic ground model to be used within the KTH model that enables accounting for the spatial variability through Markov random field theory. The new ground model employs a parametric approach to describe the properties of the Markov field, hence, enabling the simulation of the ground conditions with limited data, but does not consider the epistemic uncertainty in the model parameters. This will be the addressed in future research.

Abstract

Three-dimensional (3D) ground models enable the visualization of complex subsurface conditions in offshore wind farms, which aid engineers in understanding the spatial morphologies and interrelations of different soil layers. Due to the large areas of offshore wind farms, 3D ground models established solely based on limited geotechnical data (e.g., boreholes and cone penetration tests) might lack the required accuracy. Geophysical data, particularly seismic profile data, is capable of revealing stratigraphic information and can be obtained at a relatively low cost. This study presents a case study in which a 3D ground model for an offshore wind farm located off the Southern China coast is developed through the integration of geotechnical and geophysical data. The wind farm features complex interlayering of silty materials below the soft Holocene marine deposits due to repeated sea level changes during the Quaternary period. This created significant challenges for developing a reliable ground model. In this paper, the challenges that were faced and solutions that were applied in this project are presented and discussed.