Using metaheuristic algorithms to optimize a mixed model-based ground-motion prediction model and associated variance components

Akhani, Mohsen; Pezeshk, Shahram

doi:10.1007/s10950-022-10091-y

Cited by 7 publications

(2 citation statements)

References 53 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…A third possible solution is using machine learning in predicting site parameters and local site effect. For instance, Ji et al (2021) and Zhu et al (2022) used parameters obtained from HVSR to predict earthquake site response for stations in Japan, and Akhani and Pezeshk (2022) reduced the uncertainties in GMMs by optimizing the regression coefficients. Also, Tamhidi et al (2022) proposed a new approach in predicting ground motion IMs at a target location (uninstrumented sites), which would allow estimating the site-specific HVSR-based proxies.…”

Section: Discussionmentioning

confidence: 99%

Reducing the uncertainties in the NGA-West2 ground motion models by incorporating the frequency and amplitude of the fundamental peak of the horizontal-to-vertical spectral ratio of surface ground motions

2023

View full text Add to dashboard Cite

Ground motion models (GMMs) are developed empirically to predict ground motion intensity measures. The site effects in GMMs are usually addressed using time-averaged shear-wave velocity for the upper 30 m of the site (VS30) and shear-wave isosurface depth (i.e. Z1.0 and Z2.5). However, the former does not include the effect of layered soil structure on the site’s response, and the latter is generally inferred. Many studies have shown that site fundamental frequency can be used as an additional site proxy. Using horizontal-to-vertical spectral ratio (HVSR), one can estimate the site fundamental frequency in a fast and inexpensive way. In this article, a model is developed to incorporate site fundamental frequency and its corresponding amplification factor in the Next Generation Attenuation (NGA)-West2 GMMs to reduce the uncertainties. First, a subset of the NGA-West2 database consisting of 14,636 recordings from 298 events is selected after a two-step screening procedure and used to compute the horizontal-to-vertical response spectral ratio (HVSRPSA) of recorded surface ground motions. Second, automated methodologies previously developed by the authors are used to obtain maximum likelihood estimates of site fundamental frequency (fml), its corresponding amplitude (Aml), and associated uncertainties. Third, a mixed-effect maximum likelihood approach is implemented for residual analysis and site-term residual calculation. Finally, an HVSR-based model is developed for the NGA-West2 dataset considering the relationship between site-term residual and the HVSR-based proxies (i.e. fml and Aml). In this model, Aml is as important as fml in reducing the uncertainties. Results show that using a single HVSR-based model can consistently reduce the site-to-site variability ([Formula: see text]) for all NGA-West2 GMMs, which already include the effect of VS30, Z1.0, or Z2.5. Besides, the reduction in [Formula: see text] is period-dependent, with an average of 13% reduction. As a result of the reduction in [Formula: see text], total standard deviation ( σ) decreases by 3.5% on average.

show abstract

Section: Discussionmentioning

confidence: 99%

Reducing the uncertainties in the NGA-West2 ground motion models by incorporating the frequency and amplitude of the fundamental peak of the horizontal-to-vertical spectral ratio of surface ground motions

2023

View full text Add to dashboard Cite

show abstract

“…However, the recent development in data-driven models like machine learning models has attracted the attention of researchers to apply these advanced tools for various problems related to different engineering areas (Worden et al, 2007;Akhani et al, 2019;Sparks et al, 2020;Lange and Sippel, 2020;Zounemat et al, 2021;Gandomi et al2021;Kashani et al, 2021;Akhani and Pezeshk, 2022;Azari et al, 2022;Ali et al, 2022). Machine learning models were adopted broadly for studying the scour around hydraulic structures because traditional methods like numerical or experimental models require a lot of data, and these models are costly.…”

Section: Figure 1 An Illustration Showing Scouring Around a Bridge Piermentioning

confidence: 99%

Machine Learning Model for Estimation of Local Scour Depth around Cylindrical Bridge Piers

Ahmed¹,

Waheed²,

Ashiq³

et al. 2022

IJCE

View full text Add to dashboard Cite

Scour around bridge piers is a well-known threat to bridge stability worldwide. It can cause losses in lives and the economy, especially during floods. Therefore, an artificial intelligence approach called artificial neural network (ANN) was used to predict the scour depth around bridge piers. The ANN model was trained with laboratory data, including pier width, flow velocity, particle diameter, sediment critical velocity, flow depth, and scour depth. The data was divided into 70% for training, 15 for validation, and 15% for testing. Besides, the ANN model was trained using various training algrthins and a single hidden layer with 20 neurons in the hidden layer. The results showed that the ANN model with Bayesian regularization backpropagation training algorithm provides a better predicted scour depth with a correlation coefficient (R) equal to 0. 9692 and 0.926 for training and test stages, respectively. Besides, it showed a low mean squared error (MSE), which was 0.0034 for training and 0.0066 for the test. These results were slightly better than the ANN with Levenberg-Marquardt backpropagation with R training equals 0.9552 (MSE training = 0.0047), and R test equals 0.838 (MSE test = 0.007).On the other hand, the ANN model with a scaled conjugate gradient backpropagation training algorithm showed worse predictions (R training = 0.7407 and R test = 0.6409). Besides, the ANN model shows better outcomes than the linear regression model. Finally, the sensitivity analysis has shown that the pier width is the most crucial parameter for estimating scour depth using the ANN model.

show abstract

Multi-objective optimization of reinforced concrete cantilever retaining wall: a comparative study

Kashani

Gandomi

Azizi

et al. 2022

Struct Multidisc Optim

View full text Add to dashboard Cite

This paper investigates the performance of four multi-objective optimization algorithms, namely non-dominated sorting genetic algorithm II (NSGA-II), multi-objective particle swarm optimization (MOPSO), strength Pareto evolutionary algorithm II (SPEA2), and multi-objective multi-verse optimization (MVO), in developing an optimal reinforced concrete cantilever (RCC) retaining wall. The retaining wall design was based on two major requirements: geotechnical stability and structural strength. Optimality criteria were defined as reducing the total cost, weight, CO2 emission, etc. In this study, two sets of bi-objective strategies were considered: (1) minimum cost and maximum factor of safety, and (2) minimum weight and maximum factor of safety. The proposed method's efficiency was examined using two numerical retaining wall design examples, one with a base shear key and one without a base shear key. A sensitivity analysis was conducted on the variation of significant parameters, including backfill slope, the base soil’s friction angle, and surcharge load. Three well-known coverage set measures, diversity, and hypervolume were selected to compare the algorithms’ results, which were further assessed using basic statistical measures (i.e., min, max, standard deviation) and the Friedman test with a 95% level of confidence. The results demonstrated that NSGA-II has a higher Friedman rank in terms of coverage set for both cost-based and weight-based designs. SPEA2 and MOPSO outperformed both cost-based and weight-based solutions in terms of diversity in examples without and with the effects of a base shear key, respectively. However, based on the hypervolume measure, NSGA-II and MVO have a higher Friedman rank for examples without and with the effects of a base shear key, respectively, for both the cost-based and weight-based designs.

show abstract

Using metaheuristic algorithms to optimize a mixed model-based ground-motion prediction model and associated variance components

Cited by 7 publications

References 53 publications

Reducing the uncertainties in the NGA-West2 ground motion models by incorporating the frequency and amplitude of the fundamental peak of the horizontal-to-vertical spectral ratio of surface ground motions

Reducing the uncertainties in the NGA-West2 ground motion models by incorporating the frequency and amplitude of the fundamental peak of the horizontal-to-vertical spectral ratio of surface ground motions

Machine Learning Model for Estimation of Local Scour Depth around Cylindrical Bridge Piers

Multi-objective optimization of reinforced concrete cantilever retaining wall: a comparative study

Contact Info

Product

Resources

About