Nuova ricerca

FEDERICO PONSI

Home |


Pubblicazioni

2024 - A computationally efficient procedure for calibrating model parameters of multiple specimens [Articolo su rivista]
Vincenzi, L.; Ponsi, F.; Bassoli, E.; Buratti, N.
abstract

2023 - Mitigation of model error effects in neural network-based structural damage detection [Articolo su rivista]
Ponsi, F.; Bassoli, E.; Vincenzi, L.
abstract

This paper proposes a damage detection procedure based on neural networks that is able to account for the model error in the network training. Vibration-based damage detection procedures relied on machine learning techniques hold great promises for the identification of structural damage thanks to their efficiency even in presence of noise-corrupted data. However, it is rarely possible in the context of civil engineering to have large amount of data related to the damaged condition of a structure to train a neural network. Numerical models are then necessary to simulate damaged scenarios. However, even if a finite element model is accurately calibrated, experimental results and model predictions will never exactly match and their difference represents the model error. Being the neural network tested and trained with respect to the data generated from the numerical model, the model error can significantly compromise the effectiveness of the damage detection procedure. The paper presents a procedure aimed at mitigating the effect of model errors when using models associated to the neural network. The proposed procedure is applied to two case studies, namely a numerical case represented by a steel railway bridge and a real structure. The real case study is a steel braced frame widely adopted as a benchmark structure for structural health monitoring purposes. Although in the first case the procedure is carried out considering simulated data, we have taken into account some key aspects to make results representative of real applications, namely the stochastic modelling of measurement errors and the use of two different numerical models to account for the model error. Different networks are investigated that stand out for the preprocessing of the dynamic features given as input. Results show the importance of accounting for the model error in the network calibration to efficiently identify damage.

2022 - Bayesian and deterministic surrogate-assisted approaches for model updating of historical masonry towers [Articolo su rivista]
Ponsi, F.; Vincenzi, L.; Bassoli, E.
abstract

2022 - Impact of the model error on the neural network-based damage detection [Relazione in Atti di Convegno]
Ponsi, F.; Ghirelli, G.; Bassoli, E.; Vincenzi, L.
abstract

2022 - Surrogate-based bayesian model updating of a historical masonry tower [Relazione in Atti di Convegno]
Ponsi, F.; Bassoli, E.; Varzaneh, G. E.; Vincenzi, L.
abstract

This paper presents the surrogate-based Bayesian model updating of a historical masonry bell tower. The finite element model of the structure is updated on the basis of the modal properties experimentally identified thanks to a vibration test. In a general context, model updating results are highly affected by several uncertainties, regarding both the experimental measures and the model. Stochastic approaches to model updating, as the one based on Bayes' theorem, enable to quantify the uncertainties associated to the updated parameters and, consequently, to increase the reliability of the identification. The major drawback of Bayesian model updating is the high computational effort requested to compute the posterior distribution of parameters. For this reason, the paper proposes to integrate the classical procedure with a surrogate model. A Gaussian surrogate is employed for the approximation of the posterior distribution of parameters and the performances of the proposed method are compared to those of an Bayesian numerical method proposed in literature.

2021 - A multi-objective optimization approach for FE model updating based on a selection criterion of the preferred Pareto-optimal solution [Articolo su rivista]
Ponsi, F.; Bassoli, E.; Vincenzi, L.
abstract

Multi-objectives optimization problems are often solved constructing the Pareto front and applying a decision-making strategy to select the preferred solution among the Pareto-optimal solutions. With the aim to reduce the computational effort in multi-objective optimization problems, this paper presents a procedure for the direct evaluation of the preferred updated model, without the need to evaluate the whole Pareto front. For this purpose, the objective function to minimize is defined as the distance between a candidate point and the equilibrium point in the objective function space. The choice of the criterion of the minimum distance from the equilibrium point comes from a preliminary study carried out to assess the performances of different selection criteria. The robustness and the efficiency of the proposed procedure are assessed through the comparison with the results obtained from the estimation of the Pareto-optimal solutions and the subsequent selection of the preferred one for two numerical case studies. The proposed procedure is finally applied to the calibration of a complex FE model with respect to experimental modal data. Results show that the proposed procedure is effective and considerably reduces the computational effort. Moreover, the procedure is able to directly estimate the optimal weighting factor that allows to know the relative importance between the selected objectives and can be used to solve the multi-objective optimization with the weighed sum method.

2021 - Bayesian Model Updating and Parameter Uncertainty Analysis of a Damaged Fortress Through Dynamic Experimental Data [Relazione in Atti di Convegno]
Ponsi, Federico; Bassoli, Elisa; Vincenzi, Loris
abstract

A probabilistic analysis for the uncertainty evaluation of model parameters is of great relevance when dealing with structural damage assessment. Indeed, the identification of the damage severity associated to its uncertainty can support the decision-maker to close a bridge or a building for safety reasons. In this paper the results of the model updating of an historical masonry fortress damaged by the seismic event that hits the town of San Felice sul Panaro and the surrounding localities in the Po Valley in the 2012 are presented. A standard and a Bayesian updating procedures are first applied to the calibration of the complex Finite Element (FE) model of the fortress with respect to experimental modal data. The uncertainty of the identified parameters of structural system is then obtained by using the Bayesian probabilistic approach. The most probable parameter vector is obtained by maximizing the posterior probability density function. The robustness and the efficiency of the procedure are evaluated through the comparison with the results obtained from the estimation of the Pareto-optimal solutions.

2019 - Dynamic monitoring and evaluation of bell ringing effects for the structural assessment of a masonry bell tower [Articolo su rivista]
Vincenzi, L; Bassoli, E; Ponsi, Federico; Castagnetti, C; Mancini, F
abstract

The paper presents a rational procedure for the structural assessment of bell towers with respect to the dynamic actions induced by the bell ringing, based on structural monitoring and data processing. This methodology is tested on the bell tower of Saint Prospero (Reggio Emilia, Northern Italy), a masonry tower of great historical interest. The procedure includes an accurate geometrical survey for a deep knowledge of the structural geometry, which is the base for the structural modelling. A finite element model of the tower is calibrated with respect to the modal properties evaluated from the structural response measured in operational conditions. The structural response of the tower was also measured during a famous cultural event held in the city centre of Reggio Emilia, when skilled bell ringers played traditional melodies. The tower displacements are evaluated applying a double integration technique and a detrending procedure based on the empirical mode decomposition to the acquired accelerations. To simulate the structural response to the bell ringing, the tower and the bell are modelled as a single degree of freedom system and an unforced and undamped simple pendulum, respectively. For safety assessment purposes, the structural response of the tower is evaluated considering different oscillation angles in order to identify the one causing the maximum displacement. Finally, the stress pattern of the masonry caused by the bell-induced displacement is estimated thanks to the calibrated finite element model of the tower. Results show that the bell ringing causes a stress pattern lower than the allowable masonry strength for the serviceability conditions.

2019 - Monitoraggio di strutture storiche mediante interferometria radar terrestre [Relazione in Atti di Convegno]
Bassoli, Elisa; Ponsi, Federico; Castagnetti, Cristina; Mancini, Francesco; Vincenzi, Loris
abstract

In questa memoria vengono discusse le prestazioni, i vantaggi e gli svantaggi dell’interferometria radar terrestre con antenna ad apertura reale (TInRAR) per il monitoraggio strutturale degli edifici esistenti e storici. L’interferometro radar terrestre è in grado di rilevare spostamenti differenziali confrontando la fase del segnale radar raccolto e riflesso dalla struttura rispetto a quella del segnale trasmesso. Gli spostamenti lungo la linea di vista (LoS) che si possono misurare sono dell’ordine del centesimo di millimetro e le misure possono essere campionate con frequenza fino a 200 Hz. La memoria riporta i test dinamici effettuati sulla torre campanaria della Basilica di San Prospero a Reggio Emilia (Italia), confrontando gli spostamenti misurati dal radar terrestre con i valori ottenuti per doppia integrazione delle accelerazioni misurate da accelerometri installati direttamente sulla struttura. Grazie alla capacità di misurare in maniera continua e diretta gli spostamenti di numerosi punti della struttura, l’uso della strumentazione TInRAR integrata alla strumentazione tradizionale potrebbe essere di grande ausilio nell’interpretazione del comportamento strutturale e nella valutazione della progressione di un potenziale danneggiamento nelle strutture a seguito di eventi sismici.

2019 - Ottimizzazione multi-obiettivo di modelli a elementi finiti: criteri per la scelta della soluzione ottimale [Relazione in Atti di Convegno]
Ponsi, Federico; Bassoli, Elisa; Vincenzi, Loris
abstract

La possibilità di avere modelli ad elementi finiti accurati riveste una grande importanza nella valutazione degli effetti di un evento sismico sulla struttura e nell’identificazione di eventuali danni. Per migliorare l’accuratezza del modello numerico è possibile utilizzare tecniche di model updating per calibrare parametri fisici/strutturali del modello sulla base dei dati misurati sulla struttura reale da un sistema di monitoraggio. La maggior parte dei problemi di calibrazione di modelli di strutture reali presenta diversi obiettivi che sono generalmente in conflitto tra loro. Un approccio comune per risolvere l’ottimizzazione multi-obiettivo è quello di minimizzare una funzione a singolo obiettivo definita come la combinazione pesata dei diversi obiettivi. L’articolo presenta l’effetto sui parametri fisici/strutturali calibrati del fattore peso che governa la funzione obiettivo nel metodo della somma ponderata. Dopo aver valutato la frontiera di Pareto, ripetendo la procedura di ottimizzazione più volte per diversi valori del fattore peso, la memoria descrive e confronta i risultati ottenuti da diversi criteri per trovare la soluzione ideale tra quelle ottime che formano la frontiera di Pareto, con lo scopo di ottenere la soluzione che rappresenta quindi il miglior compromesso tra gli obiettivi diversi. Viene proposta quindi una procedura in grado di ottenere la soluzione ideale senza la necessità di calcolare l’intera frontiera di Pareto, consentendo un notevole risparmio di tempo nell’intero processo. La procedura proposta è presentata attraverso il caso studio della Rocca di San Felice sul Panaro, gravemente danneggiata durante gli eventi sismici dell’Emilia nel 2012.