Bayesian model identification of higher-order frequency response functions for structures assembled by bolted joints

Teloli, Rafael de Oliveira; Silva, Samuel da; Ritto, T.G.; Chevallier, Gaël

doi:10.1016/j.ymssp.2020.107333

Cited by 24 publications

(18 citation statements)

References 32 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Table 1 presents the lower and upper limits min and max , respectively, of the feasible solutions S. They were selected through preliminary assumptions. For instance, the observation of a softening behavior reveals the requirement of 0 <˜≤ [ 34].…”

Section: Deterministic Identification Stepmentioning

confidence: 99%

“…Although the Bouc-Wen model equations considered here to describe the BERT benchmark's response are the same as those considered by Teloli et al [34], the set of experimental realizations used for each work is different. However, a comparative analysis between both identification frameworks can be drawn, as summarized on the Tab.…”

Section: Stochastic Identification Stepmentioning

confidence: 99%

“…Considering the deterministic procedure, there are no considerable differences in the experimental tests needed once both methodologies use white noise, transient sweptsine, and sinusoidal signals in a frequency range; however, while the method based on the higher-order FRFs estimated through the Volterra series expansion considers the sequential quadratic programming as an optimization algorithm, this work uses the CE-method. Regarding the stochastic identification, the framework proposed by Teloli et al [34] identifies, firstly, the parameters˜and˜considering sine inputs with excitation frequency around one-third of the resonance frequency ( /3) to ensure a closed hysteresis loop, in which there is no dependence of the parameter˜. Depending on the mode of interest, this excitation condition may be close to another vibrating mode, causing undesirable coupling effects.…”

Section: Stochastic Identification Stepmentioning

confidence: 99%

“…Fantetti et al [11], in turn, replicated hysteresis loops from a fretting test rig using a modified Bouc-Wen model, which takes wear into account, to simulate the evolution of contact parameters. Teloli et al [34] put forward the Bouc-Wen model to fit the experimental response of the BERT benchmark, an experimental setup that consists of assembling two beams by a double bolted joint. However, compared to the Iwan model, there is a lack of contributions involving the use of the Bouc-Wen model to characterize contact interactions in bolted joints.…”

Section: Introductionmentioning

confidence: 99%

“…Starting from a priori knowledge about the structural dynamics based on the previous deterministic results, a posteriori statistical model employing the Markov Chain Monte Carlo/Metropolis-Hastings (MCMC) algorithm is identified through the maximization of the likelihood of the harmonic amplitudes obtained via HBM. In Section 4 the efficacy of this approach is demonstrated applying it on the BoltEd stRucTure (BERT ) benchmark [22,34], which is composed of two aluminum beams jointed by a symmetric double-bolted joint in a cantilever boundary condition. The preliminary deterministic model is briefly presented to illustrate the preliminary knowledge acquired and then expanded to the stochastic model using the Bayesian paradigm formulation, allowing to generate a model that can accommodate the uncertainties existing in the system through a mean and a confidence band.…”

Section: Introductionmentioning

confidence: 99%

See 4 more Smart Citations

Bayesian Model Identification Through Harmonic Balance Method For Hysteresis Prediction in Bolted Joints

Miguel

Teloli

Silva

2021

Preprint

View full text Add to dashboard Cite

Hysteresis is a nonlinear dissipative phenomenon present in many structures, such as those assembled by bolted joints. However, the approaches for identification are still limited in the literature due to their complexity. Additionally, there are also uncertainties held in bolted structures caused to fluctuations in tightening torque and pressure distribution along the contact surface. Thus, this paper yields a methodology for identifying a stochastic Bouc-Wen model for bolted joints based on the harmonic balance method. Unfortunately, some challenges are encountered when applying conventional series approximation to hysteresis, caused mainly by non-smooth behavior, which induces abrupt transitions between different motion regimes. In this work, previous adaptations were made to split the hysteresis loop in smooth paths and then use a piecewise harmonic balance approach. In this way, it was possible to deal with a deterministic Identification problem based on minimizing the error between the Fourier amplitudes of an experimental signal and those obtained through harmonic balance applying the Cross-Entropy optimization method. So, the results were extended to a stochastic model by applying the Bayesian paradigm, in which the maximized likelihood function was also based on the harmonic balance amplitudes. This methodology was demonstrated to identify Bouc-Wen parameters capable of predicting hysteresis in the BERT benchmark, composed of two aluminum beams jointed by a bolted joint in a cantilever boundary condition. Evaluating the results in the time and frequency domain and the nonlinear behavior through the hysteresis loop, it can be concluded that the method was able to identify an accurate stochastic Bouc-Wen model in predicting the dynamics of bolted structures even taking into account the probable uncertainties of the system.

show abstract

Section: Deterministic Identification Stepmentioning

confidence: 99%

Section: Stochastic Identification Stepmentioning

confidence: 99%

Section: Stochastic Identification Stepmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations