Deep Learning based Estimation of Charging Time in Solar-Powered Electric Vehicle Charging Stations Ekarna Chakraborty, Amol Chaurasia, Barnali Kundu, Suchandra Kundu, Manit Mukhopadhyay, Shantanu Datta Proceedings of the International Conference on Research in Computational Intelligence and Communication Networks Icrcicn, 2025 Proper estimation of electric vehicle (EV) charging time is crucial for optimizing station scheduling, ensuring grid stability, and enhancing renewable energy utilization. This study aims to estimate the charging duration of a solar-powered EV charging station (SEVCS) using deep learning based approaches. Key parameters such as state of charge (SOC), battery capacity, charging rate, solar irradiance, photovoltaic (PV) output, and weather conditions are considered as model inputs. Ensemble and Deep Learning (DL) models, including Random Forest (RF), XGBoost, CatBoost, LightGBM, and Multilayer Perceptron (MLP), are applied to develop robust prediction frameworks. Model performance is evaluated using different metrices Mean Absolute Error (MAE), Root Mean Squared Error (RMSE), and $\mathbf{R}^{\mathbf{2}}$ score to identify the most accurate and efficient model for charging time estimation. Experimental results demonstrate that ensemble-based deep learning models, particularly XGBoost, LightGBM, and CatBoost, achieved superior accuracy with $\mathbf{R}^{\mathbf{2}}$ values approaching 0.99. These findings confirm the effectiveness of deep learning methods in enhancing charging time prediction, supporting efficient scheduling and solar energy management in smart EV infrastructures.
Neural Network-based Techniques for Classifying IoT-Enabled Smart Irrigation Data in Agriculture Radha Mahendran, Priyanka Ravindra Dhumal, Nanthini L, Shantanu Datta, Mohaideen A, M. S. Mohamed Mallick Proceedings of the 6th International Conference on Smart Electronics and Communication Icosec 2025, 2025 Conventional farming has supported human advancement for a long time, but humanity will need to start using smarter, more effective farming techniques because of rising population constraints and water scarcity. Agriculture 4.0 has arisen in this setting, improving resource management through the use of cutting-edge technology like crop health monitoring and remote sensing. Using SIS, or Smart Irrigation Data in Agriculture, to maximise water efficiency is a crucial part of this transition. In order to enhance SIS prediction and classification, this work suggests a hybrid model called CNRN-AHB. There are two parts to the model's operation: preprocessing, which uses data refinement techniques to make sure the input is high-quality, and training, which uses a combination of CNN and RNN networks optimised by the Adaptive Honey Badger algorithm. The model's prediction capabilities are improved by the hybrid technique, which allows it to successfully distinguish between exudates and non-exudates. The CNRN-AHB model outperforms current state-of-the-art methods, according to analytical data, which demonstrate an accuracy of about 95.24 percent in classification and prediction. Contributing to sustainable farming and efficient water management within modern agriculture, this research highlights the significance of intelligent data-driven strategies in smart irrigation.
AI and Machine Learning for Energy Optimization Birudala Venkatesh Reddy, K. Anju Aravind, Mohammad Shabbir Alam, Shantanu Datta, B. Karunamoorthy, Satyajee Srivastava, V. Bhoopathy Energy Efficient Algorithms and Green Data Centers for Sustainable Computing, 2025 ML and AI can transform energy optimisation in numerous industries. This chapter discusses how AI and ML have revolutionized price, energy efficiency, and environmental sustainability. AI-powered systems can optimise the grid's renewable energy integration, manage energy resources in real time, and forecast consumption trends using optimization, and predictive analytics. Smart grids, renewable energy forecasting, industrial energy management, smart buildings, and EV charging infrastructure are major applications. This chapter also discusses these fields ML methodologies. Supervised learning estimates energy consumption, RL regulates energy adaptively, and deep learning analyzes complicated data. This chapter presents effective AI-driven energy solution case studies. Edge AI, decentralized energy management, and intelligent storage technologies are also covered. It address data security, ethical concerns, and regulatory compliance caused by AI's growing use in energy optimisation to achieve a sustainable and egalitarian future.
Development of Consensus Trust-based Mechanism with Expulsion of Malicious Nodes for Permissioned Private Blockchain Networks Nandagopal Kaliappan, Saumya Chaturvedi, R. Parvathi, T. Anuradha, M. Sathya Priya, et al. International Journal of Vehicle Structures and Systems, 2024 The main objective of this paper is to propose a consensus mechanism, based on trust, for permissioned private blockchain networks. The proposal shows how the use of the cooperativeness of the network nodes is fundamental for the development of a control system, in which the behavior of the nodes is monitored by other nodes. The monitoring of the reputation score is constant in the network and through voting, the node considered malicious to the network is expelled. For the mechanism to work, it is necessary to define rigidity criteria to be applied to actions identified as malicious what are the confidence threshold adopted and what grade receives a node that has not yet been evaluated by the network. Also, determine how many nodes are needed to monitor the behavior of a Malicious Node (MN), in order to identify and expel it more efficiently. The objectives of this work are also to present the performance evaluation of two platforms for the development of permissioned private blockchains in relation to the validation time of a transaction, the response time to chain searches and the transaction mining time and to propose a use case of blockchain technology for application in the storage of electronic medical records in a hybrid approach, which combines the application of public key infrastructure with blockchain technology, to comply with the storage requirements of medical data and offer patient-centric data privacy and access control.
DEVELOP AND TEST MAGNETIC NANO CATALYSTS REUSE FOR OZONATION IN WASTEWATER TREATMENT Oxidation Communications, 2024
Pipe network blockage detection by frequency response and genetic algorithm technique Shantanu Datta, Nitish Kumar Gautam, Shibayan Sarkar Journal of Water Supply Research and Technology Aqua, 2018 This paper deals with detection of pipeline blockage location. For this, four branched pipe network models, viz. Model 1, Model 2, Model 3 and Model 4, are considered. The first two models are considered for analytical analysis and the second two models are considered for experimental analysis. For Model 1 and Model 2, the transfer matrix method is used to develop pressure frequency diagrams. Number of peaks exceeding the threshold value is considered as a variable to find the blockage location. In Model 3 and Model 4, blockage is created by partial valve closure and periodic oscillation is created by the end valve, manually. Time domain transient pressure data are analysed by the discrete Fourier transformation technique. Afterwards, an attempt is made to establish a relationship towards detection of blockage location using a genetic algorithm. This method is applied for 10%, 20%, 30% and 40% blockage of mean pipe diameter. It is found that location of blockage is independent of number of oscillations. Pressure and velocity of fluid inside the pipeline has negligible influence towards the calculation of blockage detection. New relationships and sensitivity analysis show that blockage location is directly proportional to length of maximum straight pipeline, and square root of pressure peaks. doi: 10.2166/aqua.2018.046 s://iwaponline.com/aqua/article-pdf/67/6/543/493681/jws0670543.pdf Shantanu Datta Nitish Kumar Gautam Shibayan Sarkar (corresponding author) Department of Mechanical Engineering, Indian Institute of Technology (ISM), Dhanbad, 826004, India E-mail: shibayan.sarkar@gmail.com
