Search

Prediction of rockburst in underground engineering is of significance as it is close to the safety of support structures, personnel, and working environments. To improve the accuracy of rockburst classification, this paper proposed an ensemble classifier RF-FA model in which the random forest classifier (RF) and firefly algorithm (FA) were combined to achieve the optimum performance on rockburst prediction. Five key parameters of surrounding rock, i.e., the depth H, the maximum tangential stress σθ, the uniaxial compressive strength σc, the tensile strength σt, and the elastic energy index Wet, are selected as input variables while the rockburst intensity including none, light, moderate, and strong classes was chosen as output. A total of 279 cases worldwide were collected and used for train and test the proposed RF-FA model. The results showed that the FA can optimize the hyperparameters of RF efficiently and the optimum model exhibited a high performance on rockburst data from the independent test set and new engineering projects. The feature importance obtained by the ensemble RF-FA model indicated that the elastic energy index plays the most important role in rockburst. Besides, the proposed model showed much better accuracy on rockburst classification compared with previously existing RF models and empirical criteria, which means it is a useful and robust tool for rockburst prediction. [ABSTRACT FROM AUTHOR]

Tunnel squeezing occurs when time-dependent rock creep produces large tunnel convergence. The occurrence of tunnel squeezing may result in buget increase and time waste during tunnel construction. The aim of this study was to propose a robust classifier ensemble to predict squeezing conditions in rock tunnels. Seven individual machine learning classifiers were aggregated using weighted voting methods to establish the classifier ensemble. The weight and hyperparameters of each individual classifier were tuned using the firefly algorithm. The classifier ensemble was trained and tested on a dataset collected from published literature. Missing values in the database were replaced by various imputation methods. The results indicate that the proposed classifier ensemble achieved an accuracy of 96%, higher than that of the traditionally used individual classifiers. This robust ensemble method can be applied to other classification problems in civil engineering. [ABSTRACT FROM AUTHOR]

• A novel method was proposed for predicting permeability and unconfined compressive strength of pervious concrete. • 270 samples were prepared for building the dataset. • Permeable and mechanical properties of pervious concrete were elucidated. • Beetle antennae search was firstly used to tune the hyper-parameters of support vector regression. • The support vector regression model tuned by beetle antennae search algorithm has high prediction accuracy. Pervious concrete is a widely used construction material thanks to its good drainage characteristics. Before application, its most important properties, i.e. the permeability coefficient (PC) and 28-day unconfined compressive strength (UCS) are required to be tested. However, conducting PC and UCS tests with multiple influencing variables is time-consuming and costly. To address this issue, this paper proposed, for the first time, an evolved support vector regression (ESVR) tuned by beetle antennae search (BAS) to accurately and effectively predict the PC and UCS of pervious concrete. To prepare the dataset of the ESVR model, 270 specimens in total were prepared and casted in a controlled environment in the laboratory. The water-to-cement (w/c) ratio, aggregate-to-cement (a/c) ratio, and aggregate size were selected as the crucial influencing variables for the inputs, while PC and UCS were the outputs of this model. The results indicate that both the PC and UCS firstly increased and then decreased with increasing w/c ratio. As the a/c ratio increased, PC increased, while UCS decreased. Moreover, BAS is more reliable and efficient than random hyper-parameter selection for hyper-parameter tuning. A low root-mean-square error (RMSE) and high correlation coefficient (R) indicate a relatively high predictive capability of the proposed ESVR model. The sensitivity analysis (SA) suggests the a/c ratio and aggregate size were the most sensitive variables for UCS and PC, respectively. This pioneering work provides a simple and convenient method for evaluating PC and UCS of pervious concrete. [ABSTRACT FROM AUTHOR]

The replacement of natural coarse aggregate (NCA) with recycled coarse aggregate (RCA) in concrete mixtures offers various advantages, including conservation of natural resources, reduction of CO 2 emissions, and cost reduction. However, multiple related variables and objectives (e.g., mechanical, economic, and environmental objectives) need to be considered when optimizing mixtures of recycled aggregate concrete (RAC). This cannot be achieved through traditional laboratory- or statistics-based methods. This study proposes a hybrid intelligent system based on artificial intelligence (AI) and metaheuristic algorithms for designing optimal mixtures of RAC. To verify the proposed model, a data set containing 344 different RAC mixtures was collected from previous literature. A semi-supervised cotraining algorithm using two k -nearest neighbor (k NN) regressors with different distance metrics is developed to label the unlabeled data in the collected dataset. Different AI models are incorporated into the system for modeling the relationship between RAC strength and its influencing variables. A multi-objective optimization (MOO) model based on AI algorithms and on a multi-objective firefly algorithm is used to search for optimal mixtures of RAC. The results show that k NN-based semi-supervised cotraining can effectively exploit unlabeled data to improve the regression estimates. In the test set, A Random Forest and Backpropagation Neural Network achieve the best prediction accuracy for predicting, respectively, uniaxial compressive strength and splitting tensile strength of RAC, indicated by the highest correlation coefficients (0.9064 and 0.8387, respectively) and lowest root-mean-square errors (6.639 MPa and 0.5119 MPa, respectively). The Pareto fronts of the multi-objective mixture optimization problem are successfully obtained by the MOO model. The proposed system can also be used to optimize mixture proportions of other cementitious materials in civil engineering. • RF and BPNN achieve highest accuracy for predicting UCS (R = 0.91) and STS (0.84) of RAC, respectively. • The kNN-based semisupervised cotraining can effectively exploit unlabeled data to improve the regression estimates. • Pareto fronts of multi-objective mixture optimization problem for RAC are obtained by using AI models and MOFA. [ABSTRACT FROM AUTHOR]

• The strength of SFRC are predicted using FA-SVR; • Multi-objective firefly algorithm is proposed based on the basic firefly algorithm; • The mixture proportion of SFRC is automatically designed using a multi-objective optimization method. Steel fibre reinforced concrete (SFRC) is widely used in the construction concrete industry as it partakes an important role of evolving concrete technology. It consists of steel fibres of various shapes, sizes and geometries that influence the concrete mix composition and mechanical properties. However, compared to traditional concrete, it is difficult to design the mix proportions because more influencing variables need to be considered to optimise multiple properties including ultimate compressive strength, tensile or flexural strength and cost. Therefore, the present study proposes an artificial intelligence based multi-objective optimization model to enable an efficient method of finding the optimum mix design for SFRC. A large dataset including 299 instances for uniaxial compressive strength (UCS) test and 269 instances for flexural strength (FS) test were collected from previous literature. Support vector regression (SVR) model was applied to predict UCS and FS for SFRC. The hyper parameters of SVR models were tuned using a firefly algorithm (FA) and a sensitivity study was conducted to understand the importance of the inputs on the output variables for the algorithms. High correlation coefficients (0.91 for UCS and 0.85 for FS) were achieved on the test dataset. The FA-SVR model was then applied as the objective function for a developed multi-objective FA to search for the optimal SFRC mixture proportion. Pareto optimal solutions were obtained and served as a design guide to determine the optimal SFRC mixtures. [ABSTRACT FROM AUTHOR]

Abstract: In this paper, we present a new denoising method for the depth image of a time-of-flight (ToF) camera, based on weighted least squares (WLS) framework. The common method for ToF depth image denoising is to use bilateral filter. However, the ability of bilateral filter in edge preservation would be reduced while we attempt to smooth out larger spatial scale noise. In order to avoid this problem and preserve the edge information as much as possible, we introduce a new way to construct edge-preserving ToF depth image denoising based on WLS. We are to our knowledge the first to present a WLS-based method for ToF depth image denoising. Experimental results demonstrate that compared with bilateral filter, our proposed algorithm not only achieves better performance in edge preservation, but also improves the PSNR values of the denoised images by 0.5–2.6dB. [Copyright &y& Elsevier]

• Applicability of the Witczak 1-40D model was evaluated. • Artificial intelligence model was employed to correct the inaccuracy of the Witczak 1-40D model. • Adaptive inertia weight and the Levy flight method were combined in BAS algorithm. The applicability of the widely used Witczak model was evaluated for the first time regarding the asphalt mixtures produced in the East part of China. 16 asphalt mixtures by 2 binders and 8 aggregate gradations were designed and the dynamic modulus testing was performed under various loading temperatures and frequencies. The results obtained are similar to previous studies from different countries (areas): the dynamic modulus of the asphalt mixture in the low-temperature region is overestimated. So then, a hybrid artificial intelligence algorithm to replace the traditional Witcazak model is implemented in this paper to correct the overestimation, using the same input parameters in the Witczak model. A modified beetle antennae search (BAS) algorithm was proposed in this study to improve the searching efficiency in the random forest (RF) model. The calculation process showed fast convergence and higher efficiency. The comparative results between the predicted and actual dynamic modulus showed higher prediction accuracy at all the temperature and frequency ranges, overcoming the weaknesses of the traditional Witczak model. The variable importance results showed that G* and phase angle of the binders have the greatest impact on the dynamic modulus. Volumetric properties also show certain influence ability, but the change of the variable controlling aggregate gradation has a weak influence on the dynamic modulus. [ABSTRACT FROM AUTHOR]