Hongwei Huang
Multi-dimensional Risk Assessment and Its Application in the Construction of an undersea tunnel in China
Hongwei Huang
Professor in department of Geotechnical Engineering at Tongji University
Abstract
It is well known subsea tunnel plays a significant role in promoting regional and national economic development. However, owing to their special service environment and construction difficulty, subsea tunnel projects exhibit pronounced characteristics of highly aggregated multi-dimensional risk and strong uncertainty in risk evolution during construction. Generally, it crosses fault zones, weak fractured rock masses, and deep overburden layers, and are typically subjected to the extreme condition of “a water body overhead.” Moreover, subsea tunnels generally feature ultra-long excavation distances and often involve multi-section, multi-method construction. Shield tunnelling or TBM, drill-and-blast, and cut-and-cover methods must be precisely connected under subsea or near-shore environments. Disturbance effects induced by different construction methods, support system responses, and construction scheduling interact with one another, easily leading to spatiotemporal coupling and propagation of construction risks. Therefore, given the complex risk causation mechanisms, long action chains, and strongly coupled evolution processes in subsea tunnel construction, establishing a systematic risk assessment and dynamic control framework for multi-dimensional risk coupling has become a critical engineering challenge in subsea tunnel risk management.
In view of these high-risk characteristics in a real case undersea tunnel in China, this lecture presents multi-dimensional risk assessment and its application in this tunnel. First, the main engineering information including its main difficulties in the whole life cycle is introduced. Second, the fundamental principle, method and the coupling and cascading amplification of multiple risk losses are proposed. Then the probability prediction and application of various possible engineering hazards are presented.
For the evolutionary assessment of risk across temporal, spatial, and other dimensions, the probability density evolution method (PDEM) is adopted. Here, a generalized density evolution equation (GDEE) of multidimensional risk is proposed, as shown in Eq. (1):
For the evolution of risk with time at a fixed spatial location, the geotechnical parameters, hydrogeological conditions, construction disturbances, and support structures at that location are treated as uncertainty influence factors. The multi-dimensional probability space of the random variables is discretized and sampled. For each sample set, deterministic analysis is conducted using numerical simulation or surrogate models. The resulting risk time-history curves are then used as input data to numerically solve the GDEE of multidimensional risk. A similar approach can be employed to analyze risk evolution along other dimensions.
To account for the coupling effects among multiple risks such as safety, schedule, and investment, it is necessary to establish the joint PDF of multiple risks. Firstly, under fixed variable conditions, the marginal and joint PDF of the uncertainty influence factors affecting the aforementioned multiple risks are constructed. Subsequently, functional fitting or machine learning techniques are employed to establish association prediction models between the multiple risks and the influencing factors. On this basis, samples are drawn from the multi-dimensional probability space of the influencing factors and substituted into the multiple risk prediction models to compute the corresponding risk values, thereby reconstructing the marginal PDFs of the multiple risks. Finally, the Copula method is adopted to establish the joint PDF of the multiple risks, and the coupling effects of multiple risks on risk Xi can be quantified using Eq. (2).
Besides, this lecture focuses on dynamic quantitative prediction of construction risks in subsea tunnels. By integrating multi-source on-site hazard data, a probabilistic density evolution and posterior updating model is developed for risk variables and their influencing factors. To construct the Gaussian Copula structure, inverse normal transformation is applied to map the marginal distributions into standard normal space. On this basis, the Gaussian Copula function is established as in Eq.(3), and the joint probability density function is derived. Subsequently, the cumulative distribution function of the conditional probability density is obtained, and the Copula parameter is dynamically updated as new data are continuously incorporated. This enables dynamic prediction and quantitative evaluation of hazard occurrence intervals, providing a probabilistic representation of on-site safety management performance and offering an effective approach for risk evolution modelling and decision support.
This case study validates the effectiveness and practical value of the proposed theoretical framework in addressing complex engineering risk problems, offering a referable paradigm for the risk assessment in similar projects.
Biography
Prof. Hongwei HUANG is a Distinguished Professor of Tongji University, China. He is mainly engaged in risk assessment on Geo-structural system, underground infrastructure safety and health monitoring and inspection, etc. Currently, he is the director of Shanghai Institute of Disaster Prevention and Relief, the founding director of International Joint Research Centre for Resilient Infrastructure of Tongji University. And also, he has been the Founding Chair of Engineering Risk and Insurance Research Branch of China Civil Engineering Society since 2009. While serving as core members for international academic committees including GeoSNet, Geo-Institute on Risk Management of ASCE, TC304 of ISSMEGE, WG2 of ITA, etc., Prof. HUANG is also the Associate editor of Canadian Geotechnical Journal and ASCE-ASME Journal of Risk and Uncertainty of Engineering System, Editorial Board Members of Tunneling and Underground Space Technology, and GeoRisk. There have been numerous scientific works granted by National “973” and “863” Projects, 15 projects of National Natural Science Fund of China, and 17 major scientific research projects lead by him. And with more than 200 journal papers and more than 6 books published, over 20 keynotes in various prestigious international conferences delivered, Prof. Hongwei HUANG has also chaired more than 5 international conferences and further received 2 International Distinguished Service Awards.