When innovation and technology meet geotechnical engineering
By HKIE Geotechnical Division
1. Why do we need innovation and technology?
The Hong Kong construction industry is facing a wide range of enormous challenges, from an ageing workforce to increasing demands for productivity, site safety, quality, sustainability and competitiveness. Innovation and Technology (I&T) have been recognised as essential to overcoming these challenges and sustaining the growth of the construction industry. With these challenges in mind, the Institution’s Geotechnical Division (GE Division), the Geotechnical Engineering Office (GEO) of the Civil Engineering and Development Department (CEDD), practitioners, academia and industrial partners have partnered to foster the development of I&T for the benefit of geotechnical engineering in Hong Kong.
2. How do we implement innovation and technology?
In the 2021 Policy Address, the Chief Executive emphasised that the Government has made unprecedented strides to develop Hong Kong as an international I&T hub by investing more than HK$130 billion over four years.
“Over the past four years, more than HK$130 billion has been invested to develop Hong Kong into an international I&T hub.” - 2021 Policy Address
The GE Division has long taken an active role to facilitate collaboration among Government, practitioners, academia and industrial partners in developing I&T for the geotechnical engineering industry in Hong Kong. Despite the COVID-19 pandemic, the GE Division has achieved remarkable progress in engaging top management from different organisations to support and steer the application of innovations and new technologies in various geotechnical engineering projects. Pertinent technical webinars and e-class workshops have been held to promote I&T and assist practitioners in adapting to new challenges.
3. What innovative solutions have been or are being implemented in geotechnical engineering?
Over the past few years, the geotechnical discipline has successfully introduced and implemented numerous innovative solutions and novel technologies to the construction industry. Overall, I&T adopted can be summarised into four categories, namely, novel materials, automation and robotics, digital technology and artificial intelligence (AI).
“Novel materials, automation and robotics, digital technology and artificial intelligence – Four innovative solutions to present-day geotechnical engineering.”
4. Novel materials
4.1. Self-compacting backfill material
In practice, replacement or re-compaction of the top three metres of loose fill is commonly used for upgrading substandard fill slopes. The works are usually carried out in a pit-by-pit manner to avoid problems such as temporary slope instability and to minimise potential adverse effects to nearby existing trees. Drawbacks of this approach, however, include the need for adequate working space, labour-intensive and time-consuming placement, compaction and compliance testing of the fill.
The Nano and Advanced Materials Institute (NAMI), an institute designated by the Government as a research and development centre for nanotechnology and advanced materials, has recently developed self-compacting material for trench backfilling. The benefits of using this material, including reduced construction time and labour resources, were recognised by geotechnical practitioners. The GEO has also taken the lead to collaborate with NAMI in expanding the scope of application to slope works. Apart from a series of laboratory tests, site trials in slope upgrading works under the Landslip Prevention and Mitigation Programme (LPMitP) have also been conducted. The results indicate that the use of this novel material significantly reduces the involvement of labour and manual handling as compared to conventional fill replacement works, leading to remarkably shortened construction time.
4.2. Engineered cementitious composites
The geotechnical discipline is anticipating the development of a new type of concrete which can be bent and exhibits a certain degree of flexibility. The official name of the bendable concrete is “Engineered Cementitious Composites” (ECC), which was initially developed by a research team at the University of Michigan. It is a class of fibre-reinforced material that exhibits a remarkable degree of crack control and self-healing ability. To explore the potential application of this novel material in local geotechnical practice, the GEO, the Hong Kong University of Science and Technology and a geotechnical consultant have teamed up to adjust the composition of the ECC. The team has conducted a series of laboratory tests and site trials to shed further light on the material characteristics, site performance and long-term reliability.
In the tests and trials, fibres in ECC were discovered to alter the cracking behaviour of the concrete to favour the formation of multiple micro cracks instead of single wide cracks that commonly occur in conventional reinforced concrete or cement grout. As a result, the residual cement particles are able to further hydrate and expand in the presence of moisture, such as from groundwater, and fill up the micro cracks. This dual action has demonstrated promising corrosion resistance of the composites in the laboratory. Further enhancement of corrosion protection to steel ground reinforcement in geotechnical works such as soil nails in slopes might also occur.
5. Automation and robotics
5.1. Automated testing systems for concrete cubes and steel bars
To meet the increasing demand by the construction industry and the public for high quality laboratory testing, the GEO has recently initiated a project to develop automated concrete cube and steel bar testing systems in the public works laboratories for material compliance tests.
A key feature of the systems is that the entire testing process is automated. A fully computerised control unit is equipped in each of the systems to handle the entire automated testing process. Examples of automated processes include placing concrete cubes in curing tanks, retrieving specimens at a specified time, measuring specimen weight and dimensions, and conducting compression tests (Figure 1). The automated systems are currently being tested and will be put in use in fourth quarter of 2021 or first quarter of 2022. Apart from enhancing the efficiency of the testing services, both the quality control and safety aspects of the testing process are expected to be improved by automation.
Figure 1: Automated concrete cube testing system
5.2. Mobile robots for site inspection
The robotics industry has rapidly developed in recent years. To explore the potential use of robots in geotechnical engineering, the GEO recently procured two advanced quadruped robots, also known as “robotic dogs”, from Boston Dynamics (Figure 2). A series of performance tests is currently being conducted under various environmental settings commonly encountered in geotechnical works sites, including man-made slopes, natural terrain and underground spaces.
The quadruped robot is a four-legged robot with the potential to automatically traverse a wide variety of terrain. When compared with traditional wheeled robots, quadruped robots offer superior capabilities in terms of agility and obstacle avoidance. These robots are expected to be deployed to carry out inspections in environments which are potentially dangerous or difficult for human access.
Furthermore, the GEO has customised the quadruped robot by attaching a hand-held laser scanner, high resolution camera and an AI-based camera onto its body (Figure 2). The enhanced robot has been deployed successfully to carry out inspections of man-made slopes, landslide sites, disused tunnels and active tunnel construction sites. High-resolution images and point cloud data are simultaneously captured and used to create high quality 3D models for better visualisation of geotechnical problems and subsequent geotechnical assessment.
Figure 2: Quadruped robot customised by GEO
The GEO is now considering further expanding the application of quadruped robots in geotechnical works by attaching additional equipment and a leading-edge computing system onto the robots. Such expansions will enable high speed wireless data transmission and low-latency remote control of the robots.
6. Digital technology
6.1. Building Information Modelling (BIM)
In 2019, the geotechnical discipline established the first local BIM standards for geotechnical works. Since then, the discipline has sustained momentum by promoting and utilising BIM in various types of geotechnical engineering works. To catalyse this digitalisation process, the GEO has taken the lead in developing the digital design workflow to facilitate technological development for geotechnical works to enhance design efficiency, buildability and cost-effectiveness.
As an example, the GEO has been continuously developing digital automation for slope stability assessment on the BIM platform. Ground models comprising 3D geological and groundwater boundaries generated by digitising ground investigation (GI) data are incorporated in programmes, such as Leapfrog, to automatically generate cross-sections of ground conditions. Therefore, the output polygons can be imported directly to commonly used engineering programmes, such as SLOPE/W, for stability analysis (Figure 3).
Figure 3: BIM models for slope upgrading works and landscape works above the existing underground tunnels
6.2. Advanced numerical tools
Apart from BIM, the geotechnical discipline has been identifying other potential advanced numerical tools for geotechnical use. Such tools include the application of coupled analysis of landslide debris mobility and flexible barrier design using an advanced computer programme ‘LS-DYNA’ for automated numerical modelling. The coupled analysis is user-defined finite element formulation that allows efficient determination of dynamic responses of debris-barrier interaction and helps to develop robust and cost-effective barrier designs.
The advanced 3D graphical simulation in the coupled analysis can also be integrated into the BIM platform for slope design (Figure 4). The simulation enables designers to visualise the “completed” structure in its existing environment, modify the design layout for blending into the site settings and appraise visual impact of the proposed works. This simulation is particularly effective for public consultation and education.
Figure 4: 3D graphical simulation in the coupled analysis
6.3. Innotech Forum on Geotechnology
The geotechnical discipline has spared no effort in overcoming the geographical and physical barriers brought about by the COVID-19 pandemic as well as striving to explore continuous professional development (CPD) opportunities. The Innotech Forum, a signature event organised by the GEO with support from the GE Division, is an excellent example of this initiative.
The Innotech Forum, launched in 2018, provides a platform for industry practitioners, academics, government departments, learned societies and overseas experts to exchange knowledge and innovative ideas gained from the latest completed and ongoing projects or other technical development work related to geotechnical engineering. To cope with the pandemic situation, the 2021 forum was held virtually via a video conferencing system. The forum attracted over 1,000 local and overseas participants. Distinguished speakers from various fields in different countries shared their knowledge and work related to cutting-edge technologies, particularly in the areas of AI, robotics, advanced numerical modelling, BIM and novel construction technologies (Figure 5).
Figure 5: Innotech Forum on Geotechnology held on 26 March 2021
6.4. Virtual site visits for geotechnical practitioners
Following a suite of e-workshops and webinars, the GE Division fostered the new initiative of virtual site visits through online communication platforms, such as Zoom, Microsoft Teams and Cisco Webex. Virtual visits to various project sites have been held since October 2020. These visits included the Trunk Road T2 and Cha Kwo Ling Tunnel and Landslip Prevention and Mitigation works under the CEDD and the Relocation of the Shatin Sewage Treatment Works to Caverns under the Drainage Services Department. The potential for integrating 360-degree cameras and virtual reality (VR) technologies into the virtual site visits has been fully demonstrated, enabling users to enjoy an immersive experience. Virtual site surveillance can be conducted through 360° video livestreaming facilitated by VR goggles. These virtual site visits have attracted overwhelming interest from geotechnical practitioners and set repeated record-breaking registration rates, with over 1,000 applications received within a few days of opening enrolment for each event.
Another form of virtual experience is set to be visualised at the CEDD’s “Po Shan Drainage Tunnel - Landslide Sci-Tech Chamber”, is open to the public with guided tour available for individuals, schools or organisations. With the use of augmented reality (AR) on a mobile device, visitors to the Landslide Sci-Tech Chamber will be able to visualise how the Po Shan Drainage Tunnel was constructed by a tunnel boring machine (TBM) and how groundwater could flow into sub-vertical drains radiating away from the tunnel (Figure 6). The AR application provides visitors with an interactive experience combining reality and the computer-generated virtual environment.
Figure 6: AR experience at the Po Shan Drainage Tunnel - Landslide Sci-Tech Chamber
6.5. Digital site supervision system
A vast number of records, usually in paper form, are generated conventionally in active construction sites. The substantial manual handling and storage requirements make the paper-based record system inefficient, environmentally unfriendly, and create difficulty in monitoring, processing and retrieval. Some site records might even be lost or destroyed by site accidents or human error.
In 2020, the GEO adopted a pilot digital site supervision system in four active LPMitP works contracts, involving over 100 active construction sites around Hong Kong. An electronic Request of Inspection/Survey Check (e-RISC) Form System was developed by the Logistics and Supply Chain MultiTech R&D Centre, an institute funded by the Government, for site supervision of LPMitP works. The e-RISC Form System adopts “Blockchain” technology, which enables growing lists of records to be linked using cryptography. This technology guarantees supreme immutability, longevity of information, transparency and traceability. The e-RISC Form System has been on trial in selected LPMitP works contracts since early 2020 and has recently been modified into a cloud-based platform with further enhanced functionalities.
6.6. Remote sensing techniques
In most geotechnical projects, monitoring the actual performance of geotechnical design onsite is of paramount importance to ascertain that the works would not adversely affect the nearby facilities. With the recent rapid advancement of remote sensing technologies, the geotechnical discipline has been applying novel techniques. These include a global navigation satellite system (GNSS), light detection and ranging (LiDAR) and distributed fibre optic sensing that suit different site conditions and streamline tedious site operations. Three examples are described below.
6.6.1. Monitoring of movement using a global navigation satellite system (GNSS)
The CEDD’s Trunk Road T2 and Cha Kwo Ling Tunnel Project (T2 project) have employed a GNSS-based scheme for monitoring 3D movement of various sensitive structures. Some are the Public Works Central Laboratory (PWCL) building, the Kwun Tong Typhoon Shelter breakwater, and seawalls along the South Apron of the former Kai Tak Airport and at Cha Kwo Ling.
The GNSS involves multiple constellations of satellites, which provide signals that transmit positioning and timing data to ground-based GNSS receivers. This information enables the positions of the GNSS receivers to be determined, thereby tracking any movement induced by the construction activities. With the recent technological advancement of the Chinese Beidou Satellite System (北斗衛星導航系統) and the associated computing algorithms, remarkably accurate positioning within one to two millimetres resolution is achieved, which is commensurate with conventional engineering surveying methods. In addition, the GNSS can provide fully autonomous and real-time monitoring data at 15-minute intervals. The data is readily accessible through a dedicated web portal, which includes automatic issue of short message service (SMS) warnings and/or warning e-mails to subscribers, whenever specified “Alert, Alarm and Action” (AAA) limits are breached.
6.6.2. Territory-wide LiDAR survey and Unmanned Aerial Vehicle (UAV) photogrammetry
In 2020, the GEO conducted a second territory-wide airborne LiDAR survey to obtain detailed topographical information in Hong Kong. Compared with the first territory-wide survey carried out in 2010, the data quality acquired in the second survey was substantially improved by taking advantage of technological advancements. For instance, the density and penetration power of the laser pulses used in the 2020 survey were four times and two times higher, respectively, than those in the 2010 survey. Therefore, the 2020 LiDAR dataset has a much higher resolution and accuracy, allowing improved terrain characterisation and providing multidisciplinary support to different projects, including flood risk analysis, aviation controls and urban planning. Handheld laser scanning and UAV photogrammetry are also commonly applied in various geotechnical studies, such as surveying, capturing site conditions and generating 3D models for assessment of natural terrain landslides on remote and difficult terrain.
6.6.3. Digital rock joint discontinuity mapping
Significant advancement of digital technology and remote sensing for geological applications has occurred in recent years. This is demonstrated by the overwhelming response from practitioners who joined the benchmarking exercise on digital rock mass discontinuity survey during the Innotech Forum. In the exercise, participants utilised state-of-the-art digital data processing techniques for systematic identification, retrieval, measurement and analysis of rock mass discontinuities based on point cloud data obtained by remote sensing techniques. This exercise also provided a platform for participants to share their experience and insight on digital rock mass discontinuity mapping using different techniques. With the success of the benchmarking exercise, the geotechnical discipline is confident in advancing the application of digital technology in geotechnical assessment.
7. Artificial Intelligence (AI)
The recent breakthroughs in AI, in particular deep learning and data analytics, have led to significant transformation of almost every business sector and professional field. Having recognised the strength of AI for various image and video analytics tasks, such as image classification, object detection, image segmentation and similarity analysis, the geotechnical discipline is now identifying suitable opportunities for the technique to further enhance geotechnical practice.
Over the last few decades, the geotechnical discipline in Hong Kong has collected a huge amount and a wide range of geological, hydrogeological and geotechnical data from numerous ground investigation and construction projects. These include soil and rock type, ground stratigraphy, engineering properties of geomaterials, underground and groundwater conditions, imagery data, performance tests and site monitoring results. With such a wealth of comprehensive data, the geotechnical discipline is actively pursuing opportunities to apply AI, in particular the Convolution Neural Network (CNN) model, in geoscience and managing geotechnical risks in Hong Kong. Initial opportunities have been identified as described below.
7.1. Identification of natural terrain landslides
Under the Hong Kong slope safety system, the identification and assessment of the distribution, characteristics and hazards of natural terrain landslides have traditionally been based on manual interpretation of aerial photographs. The compilation and updating of the pertinent landslide information in the form of inventory, namely the Enhanced Natural Terrain Landslide Inventory (ENTLI), are tedious and labour-intensive. Recently, a collaboration among the GEO, consultants and academia, has been investigating the feasibility of applying change detection, feature extraction and AI techniques to automate the mapping of natural terrain landslides and determining landslide attributes from digital aerial photography and satellite images. With the aid of AI, it is anticipated that the turnaround time in updating the ENTLI can be shortened significantly from six months to one month. Furthermore, the Government’s emergency services in responding to the territory-wide and widespread landslides in the event of extreme rainfall would be greatly enhanced.
7.2. Landslide prediction
Landslide occurrence is a complicated natural phenomenon, which is affected by several dynamic and static causal factors including rainfall, geology, surface and subsurface ground and groundwater conditions, as well as slope characteristics. To facilitate the prediction of landslide frequency, statistical rainfall-landslide correlations have been established based on past landslide incidents and rainfall data recorded by the rain-gauge system maintained by the GEO and the Hong Kong Observatory. These correlations form the backbone of the territory-wide Landslip Warning System, which provides early warning messages to the Government and the community in response to landslide incidents.
The latest AI boom, powered by Nvidia’s graphical processing unit (GPU), enables machine learning algorithms to manipulate large amounts of data and instantly complete thousands of calculations in parallel. Such a high level of computing power should allow the geotechnical discipline to use advanced machine learning techniques, such as various ensemble techniques incorporating decision trees, like Random Forest and XGBoost, and artificial neural networks. These could handle the vast amount of landslide-related data and study the multi-dimensional non-linear relationships between landslide occurrence, rainfall and other potential causal factors. Similarly, the discipline also aims to use AI techniques to enhance the conventional statistical approach in developing a rainfall-based natural terrain landslide susceptibility model and frequency map. This will be used to estimate landslide propensity over Hong Kong’s natural hillsides.
8. Where do we go from here?
The GE Division is proud of the concerted efforts of the Government, our practitioners, local academia and industrial partners in applying innovative solutions and new technologies to tackle the rising challenges in geotechnical engineering. The geotechnical discipline has gained invaluable experience through application of the latest I&T, including application of novel materials, automation and robotics, digital technology and AI.
Observing our surrounding environment, it seems the world is moving at an increasingly faster pace. I&T are quickly advancing and will never stop to wait for us. To keep pace with such rapid technological advancement, it is essential for the Government, practitioners, academia and industrial partners to foster an innovative and collaborative culture in the construction industry. The GE Division together with the Government will continue to take the lead in promoting I&T in the industry by opening up geotechnical data and applying the latest technological advancements. In return, academia, practitioners and industrial partners will have many opportunities to take initiatives and make further progress with advancements. Synergy is definitely the key for success in I&T development.
As engineers in society, we have the social responsibility to apply our I&T knowledge and techniques into practice to build a smart and liveable Hong Kong. Let us rise to meet the challenges and innovate to thrive for a sustainable future!