Pouyan PressJournal of Soft Computing in Civil Engineering2588-28725420211001Machine Learning on Microstructural Chemical Maps to Classify Component Phases in Cement Pastes12013675910.22115/scce.2021.302400.1357ENEmily FordGraduate student, School of Sustainable Engineering and the Built Environment, Arizona State University, Tempe AZ 85287, USAKailasnath ManeparambilIntel Corporation, Chandler, AZ 85224, USAAdjunct Faculty, Computer Science and Engineering, Arizona State University, Tempe AZ 85287, USANarayanan NeithalathProfessor, School of Sustainable Engineering and the Built Environment, Arizona State University, Tempe AZ 85287, USAJournal Article20210830This paper implements machine learning (ML) classification algorithms on microstructural chemical maps to predict the constituent phases. Intensities of chemical species (Ca, Al, Si, etc.), and in some cases the nanomechanical properties measured at the corresponding points, form the input to the ML model, which predicts the phase label (LD or HD C-S-H, clinker etc.) belonging to that location. Artificial neural networks (ANN) and forest ensemble methods are used for classification. Confusion matrices and receiver-operator characteristic (ROC) curves are used to analyze the classification efficiency. It is shown that, for complex microstructures such as those of ultra-high performance (UHP) pastes, the classifier performs well when nanomechanical information augments the chemical intensity data. For simpler systems such as well-hydrated plain cement pastes, the classifier accurately predicts the phase label from the intensities of Ca, Al, and Si alone. The work enables fast-and-efficient phase identification and property forecasting from microstructural chemical maps.Pouyan PressJournal of Soft Computing in Civil Engineering2588-28725420211001Assessment of the Slope Stability Under Geological Conditions Using FDAHP-TOPSIS (A Case Study for Sungun Open Pit Mine)214013940610.22115/scce.2021.290413.1337ENMorteza NiromandPh.D. Student, Department of Mining, Ahar Branch, Islamic Azad University, Ahar, IranReza MikaeilAssociate Professor, Faculty of Environment, Urmia University of Technology, Urmia, Iran0000-0001-8404-3216Mehran AdvayAssistant Professor, Department of Mining, Ahar Branch, Islamic Azad University, Ahar, IranJournal Article20210613Determining the degree of slope stability is one of the most important steps in the design of open pit mines that are affected by other mining activities. So that the collapse of a part of the wall will lead to irreparable human and compensatory damages. Slope stability is affected by natural factors such as lithology, tectonic regime, rock mass conditions, climatic conditions and design factors including slope angle, slope height, pattern and blasting method. In the present study, using a combination of fuzzy approach and multi-criteria decision models, the stability and ranking of the slope stability has been investigated. For this purpose, the stability of 28 slopes of 8 large open pit mines was evaluated. In the first step of the research, after identifying the parameters affecting the slope stability and recording their values for the studied mines, the degree of importance of these parameters were determined by experts using the Fuzzy Delphi Analytical Hierarchy Process. Then the slopes were evaluated and ranked using the technique of order preference similarity to the ideal solution technique. The slope A23 with similarity index 0.742 was selected as the most desirable alternative and the slope A15 with similarity index 0.335 as the most undesirable alternative in terms of slope stability. Meanwhile, Sungun copper mine with a similarity index of 0.399 was ranked 12th in the second half of the slope stability classification table. The results showed that, the matching of research results and field observations shows the applicability of the model in the initial evaluation of slopes to determine its stability.Pouyan PressJournal of Soft Computing in Civil Engineering2588-28725420211001Effects of Window-to-Wall Ratio on Energy Consumption: Application of Numerical and ANN Approaches415613940710.22115/scce.2021.281977.1299ENAniseh SaberDepartment of Engineering, Civil, Construction and Architecture, Marche Polytechnic University, Ancona, Italy0000-0002-5576-5663Journal Article20210418Buildings account for a major part of Total Energy Consumption (TEC) in comparison to that of industry and other sections. The opening and envelope material can affect their TEC. Accordingly, this paper aims to study the effects of the window to external wall ratio (WWR) and the application of recycled panels as the building envelope on the total energy consumption in a one-floor residential building located in Iran and characterized by a semi-arid climate. To follow the sustainability criterion, we designed two concrete panels for the external walls’ envelope including a porous concrete panel and recycled ash concrete panel. The WWR varies between 5% to 95% and the optimal WWRs are separately presented for all the months. To develop the models, we used Design Builder software which its simulations are validated via field observations. For all the panels, the least energy consumption is obtained when the WWR is 5%. However, due to lighting issues, the most optimal WWR is calculated as 45-55% based on the results of the numerical simulations. Further, it is proved that the recycled ash concrete panel outperforms the porous concrete panel in terms of minimum energy consumption. Hence, it is recommended to use eco-friendly material as the external walls envelop with the WWR below 50%. The numerical simulations provided 240 data points for each panel which is exploited to develop an ANN model. The results suggested that the ANN models predict the TEC based on the month and WWR with high accuracy.Pouyan PressJournal of Soft Computing in Civil Engineering2588-28725420211001Comparison of Machine Learning Classifiers for Reducing Fitness Evaluations of Structural Optimization577314063110.22115/scce.2021.306249.1367ENTran-Hieu NguyenFaculty of Building and Industrial Constructions, Hanoi University of Civil Engineering, Hanoi, Vietnam0000-0002-1446-5859Anh-Tuan VuFaculty of Building and Industrial Constructions, Hanoi University of Civil Engineering, Hanoi, VietnamJournal Article20210923Metaheuristic algorithms have been widely used to solve structural optimization problems. Despite their powerful search capabilities, these algorithms often require a large number of fitness evaluations. Constructing a machine learning classifier to identify which individuals should be evaluated using the original fitness evaluation is a great solution to reduce the computational cost. However, there is still a lack of a thorough comparison between machine learning classifiers when integrating into the optimization process. This paper aims to evaluate the efficiencies of different classifiers in eliminating unnecessary fitness evaluations. For this purpose, the weight optimization of a double-layer grid structure comprising 200 members is used as a numerical experiment. Six machine learning classifiers selected for assessment in this study include Artificial Neural Network, Support Vector Machine, k-Nearest Neighbor, Decision Tree, Random Forest, and Adaptive Boosting. The comparison is made in terms of the optimal weight of the structure, the rejection rate as well as the computing time. Overall, it is found that the AdaBoost classifier achieves the best performance.Pouyan PressJournal of Soft Computing in Civil Engineering2588-28725420211001Application of Machine Learning Techniques in Predicting the Bearing Capacity of E-shaped Footing on Layered Sand748914063010.22115/scce.2021.303113.1360ENSafeena NazeerM.Tech. Student, Faculty of Civil Engineering, National Institute of Technology, Hamirpur, India0000-0001-5264-202XRakesh KumarDuttaProfessor, Faculty of Civil Engineering, National Institute of Technology, Hamirpur, India0000-0002-4611-9950Journal Article20210904The paper presents the prediction of bearing capacity equation of E-shaped footing subjected to a vertical concentric load and resting on layered sand using machine learning techniques and the data used in the analysis has been extracted from finite element modelling of the same footing. The input variables used in the developed neural network model were the bearing capacity of square footing, thickness ratio, friction angle ratio and the output were the bearing capacity of E-shaped footing on layered sand. Multiple layer perceptron (MLP) and multiple linear regression (MLR) prediction models were used for the determination of error metrics and the ultimate bearing capacity of E-shaped footing resting on layered sand. Finally, for the ANN model development, a model equation was developed with the assistance of weights and biases, based on the MLP and MLR model using open-source WEKA and Anaconda software respectively. Sensitivity analysis has been performed on the data sets which correlates the various input variables with the output variable of both the models. The coefficient of determination (<em>R<sup>2</sup></em>) comes out to be 0.99 and 0.98 for the MLP and MLR models respectively indicating that both the models were able to predict the bearing capacity for the E shaped footing with acceptable accuracy.Pouyan PressJournal of Soft Computing in Civil Engineering2588-28725420211001Multiple Target Machine Learning Prediction of Capacity Curves of Reinforced Concrete Shear Walls9011314063210.22115/scce.2021.314998.1381ENYicheng YangPh.D. Candidate, Civil, Construction and Environmental Engineering, Iowa State University, Iowa, United States0000-0001-9311-9243In Ho ChoAssociate Professor, Civil, Construction and Environmental Engineering, Iowa State University, Iowa, United States0000-0002-2265-9602Journal Article20211113Reinforced concrete (RC) shear wall is one of the most widely adopted earthquake-resisting structural elements. Accurate prediction of capacity curves of RC shear walls has been of significant importance since it can convey important information about progressive damage states, the degree of energy absorption, and the maximum strength. Decades-long experimental efforts of the research community established a systematic database of capacity curves, but it is still in its infancy to productively utilize the accumulated data. In the hope of adding a new dimension to earthquake engineering, this study provides a machine learning (ML) approach to predict capacity curves of the RC shear wall based on a multi-target prediction model and fundamental statistics. This paper harnesses bootstrapping for uncertainty quantification and affirms the robustness of the proposed method against erroneous data. Results and validations using more than 200 rectangular RC shear walls show a promising performance and suggest future research directions toward data- and ML-driven earthquake engineering.Pouyan PressJournal of Soft Computing in Civil Engineering2588-28725420211001Prediction of Compression Index of Marine clay Using Artificial Neural Network and Multilinear Regression Models11412414100310.22115/scce.2021.287537.1324ENRamachandiran SaisubramanianResearch Scholar, Department of Civil Engineering, Pondicherry University, IndiaV MurugaiyanProfessor, Department of Civil Engineering Pondicherry University, IndiaJournal Article20210523Compression Index (CI) is one of the frequently used soil parameters for the determination of possible settlement. In this study, the Compression Index of Marine clay is predicted using Artificial Neural network (ANN). Marine clay samples were collected from eight boreholes located at distance varying from 0.5 Km to 2.5 Km landward from the coastline of Pondicherry. The depth of boring was up to 12m. These samples were used for determining the Plastic Limit (PL), Liquid Limit (LL) and the Natural Moisture Content (NMC) and these were taken as input parameters for computing CI. These input parameters are taken as ‘data set 1’. Similar properties of soil from over 51 boreholes were considered for analysis designated as ‘Data set 2’where the depth of sampling was up to 52. These were located at a distance up to 5.0 Km from the shoreline of Puducherry distributed across the town covering a length of over 5.0 km. In Data set 2, the LL, PL, Plasticity index (PI) Specific Gravity (G), Swell Percentage, ‘N’value and the ratio of PL/LL of the soil samples were taken as input parameters for prediction of CI. The input variables were reduced in successive iterations to determine their influence in the prediction of CI. Multilinear Regression Models using the same set of inputs was compared with that of ANN. Both the analysis methods indicated that the LL and PL of soil are not only easy to determine but are competent to predict CI with a high degree of accuracy.