Loris Vincenzi

Coupling Response Surface and Differential Evolution for Parameter Identification Problems

In the present article, a new surrogate-assisted evolutionary algorithm for dynamic identification problems with unknown parameters is presented. It is based on the combination of the response surface (RS) approach (the surrogate model) with differential evolution algorithm for global search. Differential evolution (DE) is an evolutionary algorithm where N different vectors collecting the parameters of the system are chosen randomly or by adding weighted differences between vectors obtained from two populations. In the proposed algorithm (called DE-Q), the RS is introduced in the mutation operation. The new parameter vector is defined as the one minimizing the second-order polynomial function (RS), approximating the objective function. The performances in terms of speed rate are improved by introducing the second-order approximation; nevertheless, robustness of DE algorithm for global minimum search of objective function is preserved, because multiple search points are used simultaneously. Numerical examples are presented, concerning: search of the global minimum of analytical benchmark functions; parameter identification of a damaged beam; parameter identification of mechanical properties (masses and member stiffnesses) of a truss-girder steel bridge starting from frequencies and eigenvectors obtained from an experimental field test.

Differential Evolution Algorithm for Dynamic Structural Identification

In the present article, differential evolution algorithm is used to perform structural identification of mass and stiffness properties of civil structures from dynamic test results. Identification is performed initially starting from exact values of modal parameters (frequencies and mode shapes). Robustness of the algorithm is then tested by adopting pseudo-experimental input data, obtained by adding to exact data some statistic scattering, representing experimental measurement error. Different objective functions are adopted in identification procedure, and results are compared with those obtained adopting classical gradient method. The method is used to identify masses, elastic moduli, and stiffnesses of external constraints of a RC frame structure and a steel–concrete bridge. Numerical results confirm that adopting both frequencies and mode shapes instead of frequencies only strongly increases sensitivity of objective function to identification parameters. Scattering of identified parameters is much smaller, with coefficient of variation of the same order of magnitude of that of pseudo-experimental data used as input values in dynamic identification procedure.

Comparison between coupled local minimizers method and differential evolution algorithm in dynamic damage detection problems

In the present paper, a comparison is made between the Coupled Local Minimizers (CLM) method and the Differential Evolution (DE) algorithm to perform FE model updating for the damage detection in a cracked beam. CLM method is a gradient-based method with multiple local optimization runs. DE algorithm is a direct search approach which uses a population of solution vectors collecting the design parameters. Two benchmark examples of damage assessment are considered, i.e., beams under flexural vibrations with one crack and two cracks, with unknown position and depth. The effectiveness of the two methods to obtain the set of unknown parameters has been verified by performing a number of optimization processes starting from initial values of parameters selected randomly. Both exact and pseudo-experimental input data are used. A statistical analysis of the optimization results is presented. Both methods give results much better than the classical gradient optimization method. Better performances in term of speed rate and precision have been obtained by CLM when the number of identified parameters is limited. On the other hand, DE shows good efficiency when the number of parameters increases or in the case of pseudo-experimental input data.

Structural monitoring of a tower by means of MEMS-based sensing and enhanced autoregressive models

Structural Health Monitoring (SHM) methodologies are taking advantage of the development of new families of MEMS sensors and of the available network technologies. Advanced systems rely on intelligent bus-connected sensing units performing locally data filtering, elaboration and model identification. This paper describes a family of enhanced multivariate autoregressive models that can be used in SHM-oriented identification procedures and the implementation of a new advanced SHM system in the tower of the Engineering School of Bologna University. It describes also the results given by the considered procedure and a comparison of the implemented MEMS-based system with a traditional solution based on piezoelectric seismic accelerometers.

Inverse Analysis for the Calibration of FRP—Concrete Interface Law

The inverse analysis technique is used to derive a non linear mode II interface law for Fiber Reinforced Polymer (FRP) – concrete bonding starting from experimental data. The proposed interface law is based on a fractional formula and includes the non linear compliance contributions of adhesive and concrete cover at high shear stresses. It depends on three parameters (the maximum shear stress, the corresponding slip and an exponent), which are calibrated from experimental results on debonding tests. The values of maximum loads and strain profiles along the FRP plates for different bonding lengths are used. The parameter identification is performed by the inverse analysis using a Direct Search algorithm. Some considerations on the well-posedness of the inverse problem adopting different cost functions to minimize the error between experimental and numerical data are given. After the parameter identification, the numerical results obtained with the proposed interface law are in very good agreement with the experimental results.

Dynamic identification of an ancient masonry bell tower using a MEMS-based acquisition system

In this paper results of dynamic tests performed on a bell tower located in Ficarolo (Italy) are reported. After the Emilia earthquake that occurred in 2012, the bell tower reported a serious damage pattern and, as a consequence, retrofitting interventions were carried out. Dynamic tests before and after the strengthening were performed to investigate the modal properties of the bell tower and to evaluate possible changes in dynamic behavior due to the intervention. Accelerations during ambient vibrations were recorded by means of an advanced MEMS-based system, whose main features are the transmission of the data in digital form and the possibility of performing some system analyses directly on-board of the sensors. Accelerations were acquired using 11 biaxial MEMS units. First 8 modes are clearly identified, with natural frequencies in the range 0.5-9.0 Hz. Finally, a comparison between the performances of the installed MEMS-based system and a traditional analog (piezoelectric) system is carried out and results are critically compared.

DYNAMIC ANALYSES OF A CURVED CABLE-STAYED FOOTBRIDGE UNDER HUMAN INDUCED VIBRATIONS: NUMERICAL MODELS AND EXPERIMENTAL TESTS

Nowadays, pedestrian bridges are increasingly lively and slender structures due to the development of improved structural materials and aesthetic requirements. As a result of this trend, contemporary footbridges are more and more prone to human-induced vertical and lateral vibrations that can compromise the comfort serviceability conditions. The goal of this paper is to characterize the dynamic behaviour of a curved cable-stayed footbridge subjected to pedestrian loads starting from experimental tests and numerical dynamic analyses. The dynamic behaviour of the footbridge is investigated thanks to an experimental campaign performed by means of an advanced MEMS-based SHM system. Accelerations due to ambient vibrations are recorded and the modal parameters of the structure are identified by means of a classic identification method. Then, to investigate the dynamic response of the footbridge subjected to pedestrian actions, a wide number of experimental tests were performed with different-sized groups of pedestrians crossing the footbridge, running, free or synchronized walking with different pacing frequencies. Then, a finite element model of the footbridge is developed and calibrated so that the numerical dynamic predictions agree with the experimental modal properties. Then, to simulate dynamic loading conditions due to a single pedestrian or a crowd of people crossing the footbridge, two mathematical models are examined. In the first approach both the non-calibrated and the updated FE model are adopted to evaluate the vertical dynamic response of the footbridge when subjected to pedestrian loads. Dynamic analyses are performed by simulating the pedestrian walking through a periodic load model representing the human-induced force as a deterministic force. The second approach is based on the solution of the equation of motion via modal decomposition, considering multiharmonic forces and experimental mode shapes and frequencies. Finally, the accelerations obtained through the mathematical approaches are compared with the experimental results.

Structural Health Monitoring of a Historical Masonry Bell Tower Using Operational Modal Analysis

This paper addresses the modal and structural identification of the historical masonry bell tower of Ficarolo, in Italy. After the seismic sequence of May 2012, the tower reported a serious damage pattern. Retrofitting interventions were designed and they mainly consisted in the rebuilding of cracked zones and the strengthening of masonry walls with carbon bars embedded in the masonry with epoxy resin. Afterwards, a continuous dynamic monitoring system has been installed on the tower. From the recorded structural response under ambient excitation, the dynamic characteristics of the tower are identified using Operational Modal Analysis techniques. Results of the first months of continuous monitoring are presented in this paper. Moreover, in order to analyse the evolution of the structural behaviour, the effect of changing temperature on the identified natural frequencies is investigated. The experimental modal parameters are also used to identify the elastic modulus of the reinforced masonry through the calibration of a Finite Element (FE) model of the tower. In addition, the influence of the soil-foundation system on the structural behaviour is evaluated. The calibration procedure is performed adopting an improved surrogate-assisted evolutionary strategy. The calibrated FE model can be adopted to simulate the structural response to far-field earthquakes. Moreover, the monitoring system can give valuable information on the structural behaviour and the structural health in the case of seismic events.

DYNAMIC BEHAVIOUR OF THE SAN FELICE SUL PANARO FORTRESS: EXPERIMENTAL TESTS AND MODEL UPDATING

Abstract. This paper describes the experimental tests and numerical analyses performed to characterize the dynamic behaviour of the principal tower of the San Felice sul Panaro Fortress (Modena, Italy). After the Emilia earthquake that occurred in 2012, the Fortress reported serious damage, such as severe cracks on the walls and collapses of several towers and the roof. As a part of a research that aims at evaluating the vulnerability of the Fortress and designing retrofitting interventions, full-scale ambient vibration tests were performed to evaluate the dynamic properties of the principal tower. Afterwards, a Finite Element (FE) model is calibrated to obtain a good match between the numerical and experimental modal properties. The optimization process is carried out through an improved surrogate-assisted evolutionary strategy. Due to the serious damage of the Fortress, the effective stiffness of the cracked masonry and the efficiency of connection at the interface between the principal tower and the rest of the Fortress are considered the main uncertain quantities to be calibrated. A multi-objective optimization is performed, considering the frequency and mode shape residuals. These are defined as the difference between experimental and numerical modal properties. The multi-objective optimization is reduced to a series of a single-objective optimization adopting the weighted sum method. The set of optimal solutions that form the Pareto front is obtained performing the optimization for different values of the weighting factors. Then, two criteria are used and compared in order to find the preferred solution among the Pareto front solutions. Finally, a comparison of the identified structural parameters obtained varying the weighting factors for natural frequencies and mode shapes in the optimization process is presented, highlighting the importance of a proper choice of the weighting factors. Available online at www.eccomasproceedia.org Eccomas Proceedia COMPDYN (2017) 2253-2268

Human-structure interaction effects on the maximum dynamic response based on an equivalent spectral model for pedestrian-induced loading

This paper investigates the effects of the human-structure interaction (HSI) on the dynamic response based on a spectral model for vertical pedestrian-induced forces. The spectral load model proposed in literature can be applied for the vibration serviceability analysis of footbridges subjected to unrestricted pedestrian traffic as well as in crowded conditions, however, in absence of HSI phenomena. To allow for a more accurate prediction of the maximum structural response, the present study in addition accounts for the vertical mechanical interaction between pedestrians, represented by simple lumped parameter models, and the supporting structure. By applying the classic methods of linear random dynamics, the maximum dynamic response is evaluated based on the analytical expression of the spectral model of the loading and the frequency response function (FRF) of the coupled system. The most significant HSI-effect is in the increase of the effective damping ratio of the coupled system that leads to a reduction of the structural response. However, in some cases the effect of the change in the frequency of the coupled system is more significant, whereby this results into a higher structural response when the HSI-effects are accounted for.