Keywords: tunnel, knowledge representation, safety, upgrade, ontology.

In the recent literature of Artificial Intelligence many diverse applications to Civil Engineering, including Knowledge Representation methods, have been documented, that prove the capability of AI techniques to enhance meeting of Functional and Prescriptive Requirements in the processes of Designing, Performing and Managing Civil Constructions. For a general reference see [2].

However, to our knowledge, what we show here is the first attempt to apply such techniques both in the case of Upgrading Procedures and for long tunnels. We argue, in the rest of this paper, that these two specificities are relevant for various reasons.

First of all, in upgrade procedures of pre-existent constructions we generally do not have data to manage, since the original data used to control Project are not in our hands, and typically are not available at all. This is a disadvantage in the sense that we often need to build up models from scratch, but allows us to provide less ambiguous modeling. This means that a significant attention has to paid to the meta-level where we have knowledge representation instead of Data Handling.

Secondly, long tunnels are peculiar in the panorama of civil engineering, with respect to meeting functional and prescriptive requirements, since functional and prescriptive requirements of long tunnels are not only constraints on functionalities, but definite limits to the factual existence of systems; a tunnels that does not meet the safety requirements for rescue is simply out of use.

Moreover, Knowledge Representation for civil engineering is not generally centered on requirements, since data handling relies on project information, which is typically much more procedural than functional;

Again, AI techniques are often used for performing in a quicker fashion standard tasks; the use of knowledge representation methods provide the opportunity of performing (often in an inefficient way) non-standard tasks; the challenge of upgrade procedure for long tunnels is that the required performance needs both the ability of solving non-standard tasks and to perform them efficiently (in real-time);

Last, but not least, the problems involved in reasoning about functional and prescriptive requirements for long tunnels is definitely highly differentiated, since categories of functional and prescriptive requirements are many, and very differently characterized; we undoubtedly need methods for organizing data handling more efficiently and systematically than with traditional applications of knowledge representation to civil engineering. In particular, we need the support of novel applications of knowledge representation to real-world domains, which recur on formal ontology.

The application we shall describe in this paper will result from an activity of a work package in a project called UPTUN (Upgraded Tunnels), whose purpose is the development of a procedure for upgrading existing tunnels (for a reference see [1]. The activity consists in fact in the development of a system that helps the experts involved in the upgrading process of long tunnels in managing the knowledge used for making the decisions that implement the upgrade procedure.

1

G.A. Khoury, EU tunnel fire safety action, Tunnels & Tunnelling International, April 2003.

2

B. Kumar and B.H.V. Topping, Artificial Intelligence Applications in Civil and Structural Engineering, Civil-Comp Press, Edinburgh, 1999. doi:10.4203/ccp.62

purchase the full-text of this paper (price £20)

go to the previous paper
go to the next paper
return to the table of contents
return to the book description
purchase this book (price £82 +P&P)