RT - Journal Article
T1 - Estimation of Lateral Load Capacity of Piles Using a New Intelligent Combination Method
JF - mdrsjrns
YR - 2022
JO - mdrsjrns
VO - 22
IS - 5
UR - http://mcej.modares.ac.ir/article-16-53735-en.html
SP - 223
EP - 233
K1 - Relevant vector regression
K1 - Artificial bee colony algorithm
K1 - Lateral load capacity
K1 - Sensitivity analysis
AB - Estimation of the load carrying capacity of pile foundation is one of the most sought after research areas in geotechnical engineering. Static equilibrium and other dynamic equations are used to predict the axial load capacity of pile. The prediction of lateral load capacity of piles, used in tall and offshore structures is more complex and requires solution of non-linear differential equations. The elastic analysis adopting Winkler soil model is not suitable for the non-linear soil behavior. Estimating the load capacity of such piles using experimental methods is always associated with error and makes the modeling result far from reality. Today, intelligent methods have shown a high capability in predicting and estimating unknown variables and can replace experimental and analytical methods. In this research, we tried to accurately predict the lateral load capacity of piles in clay soils by creating an intelligent hybrid model called optimized relevant vector regression with the artificial bee colony algorithm. The relevant vector regression is a probabilistic method based on Bayesian approach. The relevant vector regression does not need to predict the error/margin tradeoff parameter C, which can decrease the time and the kernel function, does not need to satisfy the Mercer condition. For those relevant vector regression advantages compared with the support vector regression approach, relevant vector regression model is successfully applied in regression prediction problems. In this method, relevant vector regression is used as a predictive model and artificial bee colony algorithm is used to optimize the parameters of relevant vector regression method. The artificial bee colony algorithm is a swarm based meta-heuristic algorithm for optimizing numerical problems. It was inspired by the intelligent foraging behavior of honey bees. The algorithm is specifically based on the model for the foraging behavior of honey bee colonies. The model consists of three essential components: employed and unemployed foraging bees, and food sources. The first two components, employed and unemployed foraging bees, search for rich food sources, which is the third component, close to their hive. The model also defines two leading modes of behavior which are necessary for self-organizing and collective intelligence: recruitment of foragers to rich food sources resulting in positive feedback and abandonment of poor sources by foragers causing negative feedback. In artificial bee colony, a colony of artificial forager bees (agents) search for rich artificial food sources (good solutions for a given problem). To apply artificial bee colony, the considered optimization problem is first converted to the problem of finding the best parameter vector which minimizes an objective function. Then, the artificial bees randomly discover a population of initial solution vectors and then iteratively improve them by employing the strategies: moving towards better solutions by means of a neighbor search mechanism while abandoning poor solutions. In this modeling, the data used are related to a laboratory data set of small-scale pile load capacity. Various statistical indicators were used to evaluate the modeling accuracy. Finally, the results showed that the combined relevant vector regression with the artificial bee colony algorithm for test data with R2 = 0.975 and RMSE = 0.001, has a high ability to predict the lateral load capacity of spark plugs. In addition, the sensitivity analysis performed in this study showed that the variables of eccentricity of load and the length of pile are more important and effective compared to other parameters.
LA eng
UL http://mcej.modares.ac.ir/article-16-53735-en.html
M3 10.22034/22.5.223
ER -