Faculty Publications
Permanent URI for this communityhttps://idr.nitk.ac.in/handle/123456789/18736
Publications by NITK Faculty
Browse
6 results
Search Results
Item Crack Detection in Concrete Using Artificial Neural Networks(Springer Science and Business Media Deutschland GmbH, 2023) Palanisamy, T.; Shakya, R.; Nalla, S.; Prakhya, S.S.This paper aims to explore the possibility of using machine learning (ML) algorithms and image processing to determine cracks in concrete and classify them as Cracked and Uncracked. This is a very current field of study with a lot of research currently taking place. In particular, neural network algorithms such as VGG16, ResNet50, Xception and MobileNet, were used to name a few. Two datasets were used to detect the presence of cracks in concrete. The first two datasets were taken from the Kaggle website. The first dataset is generated from 458 high-resolution images (4032 × 3024 pixels). This dataset consists of 40,000 images, 20,000 with and 20,000 without cracks. The second dataset had pictures of cracked and uncracked decks on a bridge from a dataset called SDNET2018 (2018). VGG16 Architecture based artificial neural network performed the best while MobileNet performed the worst. As the scope for the project expanded, an effort was made to determine crack properties, specifically crack width as an automated system for the same would be much more useful than a manual one. It was done using morphological transformations and concepts of Euclidean distance. © 2023, The Author(s), under exclusive license to Springer Nature Switzerland AG.Item Adaptive Neuro-Fuzzy Systems and Ensemble Methods in Joint Shear Prediction and Sensitivity Analysis(Springer Science and Business Media Deutschland GmbH, 2024) Palkar, S.S.; Palanisamy, T.In the absence of ductile design, beam-column joints form weak links in the frame during seismic activities, hence jeopardizing the entire structure. Deducing from the views of researchers, estimation of joint shear strength of RC beam-column joint is a necessity with a complexity. This complexity highlights the importance of machine learning models due to their data handling and predictive capabilities. This study used 233 beam-column joints with 132 exterior and 101 interior joints for training and testing the ensemble machine-learning models and an Adaptive neuro-fuzzy inference system. The performance indices of the models built and their comparison is carried out to find the optimum model to be deployed. The sensitivity analysis of the features considered was conducted to infer the differences in exterior and interior beam-column joints’ behavior. © 2023, Springer Science and Business Media Deutschland GmbH. All rights reserved.Item Prediction of Pore Solution Concentration in Cement Composite System by Using Machine Learning Techniques(Springer Science and Business Media Deutschland GmbH, 2024) Walke, S.; Sundaramoorthi, S.; Palanisamy, T.A thorough understanding of the pore solution's composition is crucial for a number of cementitious material properties, including durability. The pore solution concentration is determined by a variety of experimental techniques. However, these approaches aren't always straightforward. A possible substitute to complex pore solution extraction and analysis procedures could be machine learning (ML) models. The objective of this research is to explore ML techniques for predicting the cement pore solution composition composite systems produced with Ordinary Portland cement (OPC) and supplemental cementitious materials (SCM). Data on the compositions of pore solutions for different cementitious systems were gathered from the literature and combined into a comprehensive database that has over 400 data entries. Random Forest and Gradient Boosting techniques were applied to the database. Statistic metrics such as R2, RMSE and MAE were used to evaluate the prediction accuracy of the built model. Sensitivity analysis of the built models was carried out and compared. The gradient boosting technique was found to be the most effective method in prediction of the pore solution concentration (R2 ranging from (0.80–0.98) and lower RMSE values) due to its effective problem-solving capacity and minimum requirement for future engineering. Thus, ML models offer a potential approach for determining the pore solution concentration. © The Author(s), under exclusive license to Springer Nature Switzerland AG 2024.Item Improving Structural Safety with Machine Learning: Shear Strength Prediction in Interior Beam-Column Joints(Institute of Electrical and Electronics Engineers Inc., 2024) Sidvilasini, S.; Palanisamy, T.Determining the shear capacity of joints between columns and beams within a structure is crucial to guarantee its safety and stability. It directly impacts buildings' structural integrity, cost-effectiveness, and resilience, making it a critical aspect of structural engineering and construction. Estimating shear properties in beam-column joints is done via machine learning due to its ability to capture complex relationships, adapt to diverse data, and automatically identify relevant features, potentially offering improved accuracy and insights compared to traditional methods. This paper includes creating a machine-learning regression model for predicting joint shear strength in interior beam-column joints. It involves the analysis of a comprehensive dataset comprising 445 data points with 17 variables sourced from 100 research papers. The primary objective is to craft a machine-learning regression model capable of accurately forecasting joint shear strength. To achieve this goal, a multitude of methodologies have been explored, including the application of 2 machine learning regression techniques and two codes of practice (Step-wise Linear Regression, Medium neural networks, and EN 1998-1:2004, NZS 3101:1-2006). Of the two methods, step-wise linear regression gave the best results in predicting the shear capacity of interior column-beam connections. © 2024 IEEE.Item Neural network prediction of joint shear strength of exterior beam-column joint(Elsevier Ltd, 2022) Alagundi, S.; Palanisamy, T.Beam-Column joints are the critical locations in the reinforced concrete structures as they experience a massive amount of deformations under earthquake. The shear failure of the beam column joint should be avoided for the safety of the structure. In the present study, prediction of joint shear capacity of exterior Beam-column joint is proposed using artificial neural network (ANN). Experimental investigations performed by different authors have been examined and used to prepare the data sets for training, testing and validating the neural network. Parameters responsible for the shear strength of the exterior Beam-Column Joints are identified and the artificial neural network model is proposed to predict the joint shear strength. Input parameters for the ANN model are width and depth of the joint, concrete compressive strength, length of beam, top and bottom longitudinal reinforcement in the beam, yield strength of longitudinal reinforcement in beam, ratio of beam to column depth, joint Shear reinforcement index, beam bar index and column load index. The performance of the neural network model is evaluated by the statistical relations like Coefficient of correlation, Root mean square error and Scatter index. The proposed model is compared with an empirical formula and different equations suggested by the design codes. The results show that the proposed neural network model can effectively predict the joint shear strength of the Exterior Beam-Column joint. © 2022 Institution of Structural EngineersItem A Hybrid Machine Learning Approach for Predicting Joint Shear Capacity in Beam-Column Connections(Springer Science and Business Media Deutschland GmbH, 2025) Sidvilasini, S.; Palanisamy, T.Accurately predicting the shear strength of beam-column connections is crucial for maintaining the structural integrity and stability of buildings, especially in seismic conditions. This study aims to address this challenge by developing and evaluating multiple machine learning regression models for estimating joint shear capacity. A dataset consisting of 445 beam-column connections with 17 key influencing variables was compiled and used to train seven distinct regression models. Among them, the four best-performing models—Quadratic Support Vector Machine (QSVM), Rational Quadratic Gaussian Process Regression (RQGPR), Kernel Ridge Regression (KRR), and Ensemble Boosting (EB)—were selected based on their predictive accuracy. To further enhance performance, these models were combined into a hybrid ensemble model, capitalizing on their complementary strengths to improve shear strength estimation. The hybrid model exhibited superior predictive performance, achieving a test RMSE of 0.0246 and an R2 value of 0.9605, significantly surpassing the accuracy of the best standalone model (RQGPR). This reinforces the advantage of ensemble learning in minimizing error and enhancing generalization. The findings of this research highlight the growing role of machine learning in structural engineering, particularly in advancing shear strength prediction methodologies. By demonstrating that a hybrid model can outperform traditional single-model approaches, this study provides valuable insights for developing safer, more resilient structures and optimizing modern engineering practices with artificial intelligence. © The Author(s), under exclusive licence to Shiraz University 2025.