DOI:N/A

Extracted Abstract:

Upon completion of bridge inspection, monitoring and analysis, the subse- quent and crucial step is to formulate decisions for bridge maintenance, which should be logic or evidence based, optimal considering economic and environmental constraints, and feasible subject to resources at hand. There exists a situation that the decision-making is straightforward if the data and information are of low complexity. The reality is often the opposite, that is, data and information from the life-cycle inspection and monitoring processes, when brought to the maintenance engineers and managers, are often non-structured, high-dimensional, mutually dependent, and still ‘big’ in their volume. Recognizing these challenges, this chapter focuses digital tech- nologies that can facilitate scientific and rational decision-making for bridge mainte- nance and management. Among them, Artificial intelligence (AI) technologies can facilitate decision-making taking the role of assisting human-based decision-making or of driving the decision-making. Two human–machine interfacing technologies are introduced, including virtual reality (VR) and augmented reality (AR), which provide advanced visual analytics for facilitate spatial–temporal understanding, logic formu- lation, and rational decision-making. General applications and specific applications of these technologies in the context of bridge maintenance and management are highlighted too in this chapter. 9.1 AI Based Decision-Making 9.1.1 Overview Currently, bridge maintenance strategies primarily rely on bridge inspection results and reference to corresponding maintenance codes, such as the “Code for Main- tenance of Highway Bridges and Culverts” (JTG 5120—2021), while taking into account the professional expertise and experience of the decision-makers. For instance, when dealing with crack damage in the substructure of a bridge, the bridge maintenance decision-maker may formulate a maintenance decision based on recom- mendations in the code and the results of on-site crack inspections, combined with © China Communications Press Co., Ltd. 2024 G. Wu et al., Intelligent Bridge Maintenance and Management, Springer Tracts in Civil Engineering, https://doi.org/10.1007/978-981-97-3827-4_9 403 4049 Decision-Making Technologies for Intelligent Maintenance ... their own maintenance experiences. This could involve counter-measures such as applying epoxy resin sealant or pressure grouting. It is apparent that this process is filled with subjective uncertainties. Often, different decision-makers might propose different maintenance strategies for the same situation. This implies that the maintenance strategies formulated by different decision-makers will have variations in quality, leading to differences in reliability, remaining service life and other indicators of the bridge health after maintenance. In this context, if the subjective uncertainties introduced by the decision-makers them- selves in the current decision-making process are not addressed properly, even with the use of various advanced technologies to address the issues in bridge inspection, there is no guarantee that the bridge, based on the current maintenance strategy, will continue to meet the requirements of normal use, ultimate load-carrying capacity and serviceability in a satisfactory manner. Therefore, the primary goal of bridge mainte- nance decision-making is to achieve a structured and standardized decision-making process, reducing subjectivity and ensuring quality of bridge maintenance. Meanwhile, due to the existence of various objective factors, such as inability of AI technology to completely replace human involvement and inherent random- ness in the external environment where bridges are located, the maintenance of different bridges still relies on the decision-makers’ experiences. Therefore, formu- lation of bridge maintenance strategies cannot completely exclude involvement of the decision-makers. However, in order to better develop a system for formulating bridge maintenance strategies, it is necessary to track the effectiveness of the final maintenance decisions for each bridge, quantitatively evaluate the decision-making level based on the assessment results of post-maintenance inspections, and compile them into a database. This will facilitate continuous improvements in the later stages. In addition, the current formulation of bridge maintenance strategies usually focuses on individual bridges, without considering the overall management, coordi- nated planning and formulation of maintenance strategies for other bridges within the same road network. As a result, bridge maintenance often requires long duration and frequent traffic closures, leading to a waste of time and money. In summary, we need to reform the current bridge maintenance decision-making process to introduce more structured, standardized and objective decision-making methods and systems, and eliminate the subjective uncertainties introduced by the decision-makers themselves. At the same time, in the decision-making process, we should shift from considering individual bridges to considering the overall road network where the bridges are located. This allows for coordinated planning, avoiding unnecessary waste, optimizing resource allocation, extending the lifespan of bridge clusters, and ensuring the safety of people’s lives and properties. Additionally, it is necessary to establish a reasonable evaluation system to quantitatively assess the bridge maintenance strategies formulated for each bridge. This will facilitate further improvement of the bridge maintenance strategy formulation system in the future. With this background, coupled with the advancement of computer science, the concept of AI based decision-making has been proposed to overcome the current deficiencies in bridge maintenance by incorporating AI technology into the bridge maintenance decision-making process. 9.1 AI Based Decision-Making405 AI based decision-making, as the name suggests, refers to the