^ SpringerWienNewYork CISM COURSES AND LECTURES Series Editors: The Rectors Manuel Garcia Velarde - Madrid Jean Salen^on - Palaiseau Wilhelm Schneider - Wien The Secretary General Bemhard Schrefler - Padua Executive Editor Carlo Tasso - Udine The series presents lecture notes, monographs, edited works and proceedings in the field of Mechanics, Engineering, Computer Science and Applied Mathematics. Purpose of the series is to make known in the international scientific and technical community results obtained in some of the activities organized by CISM, the International Centre for Mechanical Sciences. INTERNATIONAL CENTRE FOR MECHANICAL SCIENCES COURSES AND LECTURES - No. 469 PARAMETER IDENTIFICATION OF MATERIALS AND STRUCTURES EDITED BY ZENON MROZ TECHNICAL UNIVERSITY OF WARSAW, POLAND GEORGIOS E. STAVROULAKIS UNIVERSITY OF lOANNINA, GREECE AND TECHNICAL UNIVERSITY OF BRAUNSCHWEIG, GERMANY SpringerWien NewYork The publication of this volume was co-sponsored and co-financed by the UNESCO Venice Office - Regional Bureau for Science in Europe (ROSTE) and its content corresponds to a CISM Advanced Course supported by the same UNESCO Regional Bureau. This volume contains 180 illustrations This work is subject to copyright. All rights are reserved, whether the whole or part of the material is concerned specifically those of translation, reprinting, re-use of illustrations, broadcasting, reproduction by photocopying machine or similar means, and storage in data banks. © 2005 by CISM, Udine Printed in Italy SPIN 11583752 In order to make this volume available as economically and as rapidly as possible the authors' typescripts have been reproduced in their original forms. This method unfortunately has its typographical limitations but it is hoped that they in no way distract the reader. ISBN-10 3-211-30151-8 SpringerWienNewYork ISBN-13 978-3-211-30151-7 SpringerWienNewYork PREFACE The nature and the human creations are full of complex phenomena, which sometimes can be observed but rarely follow our hypotheses. The best we can do is to build a parametric model and then try to adjust the unknown parameters based on the available observations. This topic, called parameter identification, is discussed in this book for materials and structures. The present volume of lecture notes follows a very successful advanced school, which we had the honor to coordinate in Udine, October 6-10, 2003. The authors of this volume present a wide spectrum of theories, methods and applications related to inverse and parameter identification problems. We thank the invited lecturers and the authors of this book for their contributions, the participants of the course for their active participation and the interesting discussions as well as the people of CISMfor their hospitality and their well-known professional help. Zenon Mroz Georgios E. Stavroulakis CONTENTS Preface An overview of enhanced modal identification by L. Bolognini 1 The reciprocity gap functional for identifying defects and cracks by H.D. Bui, A. Constantinescu and H. Maigre 17 Some innovative industrial prospects centered on inverse analyses by G. Maier, M. Bocciarelli andR. Fedele 55 Identification of damage in beam and plate structures using parameter dependent modal changes and thermographic methods by Z. Mroz andK. Dems 95 Crack and flaw identification in statics and dynamics, using filter algorithms and soft computing by G.E, Stavroulakis, M. Engelhardt andH. Antes 139 Application of advanced optimization techniques to parameter and damage identification problems by V. Toropov and F. Yoshida 177 Neural networks in the identification analysis of structural mechanics problems by Z. Waszczyszyn and L. Ziemianski 265 An Overview of Enhanced Modal Identification Luca Bolognini RSI Sistemi s.p.a., Altran Group, Milan, Italy Abstract. The application of identification techniques is spreading in many engineering fields, as a rehable and reaUstic understanding of structures is becoming a common issue. The aim of this paper is to summarize the main steps needed for a complete and up-to-date identification process, referring to the modal approach. The complex eigenvalue approach and a non-proportional damping model are discussed in detail. 1 Why Identification The construction industry is increasingly involved in the survey, assessment, alteration or refurbishment of existing structures. This is due to limited space and funds for new construction, the disruption costs, a greater environmental awareness and a cultural desire to maintain ancient and historic structures. At the same time, managing and operation of those structures having high social and economical impact (such as bridges, dams, commercial facilities) are demanding reliable condition-based safety evaluations. Theoretical models of civil engineering structures are essential for assessment and for the design of alterations. For older structures these models often do not exist. Where they do exist (mostly based on visual survey) they are often inaccurate, and anyway "as-built" information often does not accurately reflect the work done on site. Given the shortcomings in design data and surveys as a basis for structural models, it is clear that better non-destructive methods are needed for creating reliable models that accurately reflect the form and condition of existing structures. This process of adapting theoretical, approximate models to accord with observed behaviour is called structural identification. It can be used for the most general structural component (in both mechanical and civil fields) with the main purpose of: improving the design stage (virtual modelling); enhancing the safety assessments; - monitoring the evolution of structural conditions; - evaluating the effect of modifications or refurbishment The modal approach (aiming to the identification of structural intrinsic properties such as frequencies, modal deformations, damping factors and non linear behaviours) is one of the possible techniques that can be used for the purpose. Many efforts have been spent to make this kind of identification general, reliable and easy to manage. In the following, a short excursus in the up-to-date modal identification practice will be given, with a few more detail as far as the complex approach and a non-proportinal damping model are concerned. L. Bolognini 2 The Identification Process The identification basic idea is to improve our knowledege on a given structure through a model that accords with observed behaviour. In other words, to meet the real structure with its virtual modelling. The main steps of a complete identification process are summarized below. - Starting from the Real Structure (Step 1), - a Finite Element Model is defined (Step 2). - To understand the structure's dynamic behaviour some Prehminary Modal Analyses are performed (Step 3) and - the sensors lay-out is optimised by means of a Data Acquisition Planning (Step 4). - It is then possible to get the experimental data, Testing the Structure (Step 5). - Raw data are processed performing the Experimental Modal Analysis (Step 6), to derive the dymanic features of the real structures, which are compared with those obtained from the Finite Element model. - The discrepancies are minimised modifying the material parameter values via the Modal Matching / Model Updating iterative procedure (Step 7). The identification loop aims to a better understanding of the structure's current state and a more realistic model. In this view, many of the concepts hold for any other approach, substituting the modal one. 2.1 The Real Structure The starting point of any identification process is the structure as such (Step 1). The safety assessment of a given structure starts from a correct interpretation of its actual status. Figure 1. A typical example of structure to be identified: a masonry arch bridge. The geometrical shape can be correctly captured, different cortical construction materials can be mapped by visual inspection and boreholes can be used for fillings. An Overview of Enhanced Modal Identification The behaviour of any structure is mainly a function of: the geometrical shape; - the materials it is made of; - the boundary conditions; - the forces that are applied on it. A good approximation of the structural shape and an estimate of materials distribution are generally achievable (Figure 1). Visual inspection, boreholes, and any available documentation are useful for the purpose. The applied forces depend on the structure usage and can be prescribed, measured or controlled. From the previous considerations, material parameters and boundary conditions are the most uncertain properties and may vary in time. It is the purpose of an identification tool to identify these properties. Once that the process has been completed, more reliable stress and safety analyses will follow. 2.2 The Finite Element Model The process is based on the Finite Element model of the structure under consideration (Step 2). The geometry of the structure must be described in terms of a mesh (Figure 2). Critical issues are: the respect of the actual shape; - an adequate mesh subdivision: usual validation criteria available in most mesh modellers can be used; an alternative is to subdivide into 5-10 elements the shortest wave length involved in the modal analysis; a proper definition of element groups sharing the same material properties, which will be involved in the model updating procedure; expected material distribution, potential fault locations as well as any location of interest to the engineer should be included; - a realistic simulation of the boundary conditions applied on the structure. Figure 2. The 3D mesh of the masonry arch bridge used for identification: different material properties and boundary conditions are assigned.