Prediction of accidental actions likely to occur in building structures. An approach based on stochastic simulation (1).

Norkus, Arnoldas

PREDICTION OF ACCIDENTAL ACTIONS LIKELY TO OCCUR IN BUILDING STRUCTURES. AN APPROACH BASED ON STOCHASTIC SIMULATION (1)

Egidijus R. Vaidogas

The recently published book deals with problems of structural design for accidental actions. It has been written primarily to provide information on predicting accidental actions and assessing damage from them by means of a stochastic simulation. Both prediction and damage assessment are based on methods of the quantitative risk assessment.

[ILLUSTRATION OMITTED]

The phenomenon of accidental actions has been under investigation for a long time and the number of publications on this subject is considerable. Publications related to accidental actions are highly fragmented and insufficiently specific. One can state that there are a lot of general discussions on accidental actions. However, the number of practical recipes, how to predict them by keeping in mind the complicated problems of uncertainty and data scarcity, remain few in number.

The aim of this book is to provide a discussion on specific methodological aspects of predicting accidental actions in the design of the structures which can be subjected to them. This discussion seeks to systematise methodological tools potentially suitable for solving the problems of uncertainty and data scarcity which are almost inseparable attributes of accidental actions. The main idea of the book was to apply extensively methods of quantitative risk assessment to predicting accidental actions. These methods are not a panacea capable of "curing all illnesses" related to accidental actions; however, they allow to deal with limited information on accidental actions in a consistent way.

The book suggests computational procedures for choosing mathematical models of accidental actions and/or assessing damage to structures due to these actions. These procedures are focussed on quantifying and propagating uncertainties related to accidental actions. The key methodological tool used for propagating uncertainties was a stochastic (Monte Carlo) simulation. An application of the stochastic simulation is considered throughout the entire text of the book. Therefore the term "stochastic simulation" was included into its subtitle.

The author decided to restrict himself to the problem of how to predict actions. A detailed investigation of the behaviour of structures subjected to accidental actions was not the aim of the book. Methods of structural analysis for accidental actions are sufficiently well-developed in the deterministic structural dynamics and structural reliability theory. However, these methods can be applied effectively only when a "tenable" probabilistic model has been chosen for the accidental action in question. In this book, the structural part of the problem is "hidden" behind the so-called fragility functions. Along with a probabilistic action model, they are key elements which are indispensable to assessing the hazard of particular damage from accidental actions.

The author believes that risk-based prediction of accidental actions and potential damage from them outperforms its deterministic counterpart. Hopefully, the deterministic approach will, sooner or later, be replaced by a much more natural probabilistic, risk-oriented approach to dealing with accidental actions. The author hopes that his book will make at least a small contribution to this.

(1) E. R. Vaidogas. Prediction of accidental actions likely to occur in building structures. An approach based on stochastic simulation. Vilnius: VGTU Publishing House: Technika, 2007. 248 p.

Prof Dr Arnoldas Norkus

Vilnius Gediminas Technical University