Gaya College of Engineering

Thunderstorms are a meteorological phenomenon associated with lightning, winds, and the risk of heavy rains. Sometimes, atmospheric conditions are favorable for such dangerous thunderstorms. These severe thunderstorms may lead to great potential casualties to human life, agriculture, and property. Therefore, many researchers and scientists are working in the area of the prediction of thunderstorms. Machine learning methods have been implemented in various fields of predictive research, and also suitably in meteorology. In this paper, different machine learning techniques have been used on a localized meteorological dataset for the city of Ranchi, India, to classify thunderstorm days and others. Different DM and SC methods, namely, Gradient Boosting (GB), Naïve Bayes (NB), AdaBoost (AB), Multi-Layer Perceptron Neural Network (MLPNN) and Random Forest (RF) with different base learners such as Logistic Regression (LR) have been applied on average meteorological data of ten years (2008-2017) and hourly data of three years (2016 to 2018) in the classification of thunderstorm days. Based on this study, the boosting algorithm gave the best F-Score among all applied methods for the average dataset. However, the interest class of prediction (thunderous day) is low due to the imbalance of the dataset. Thus, some cases of non-thunderstorm days have been removed to get a modified average dataset. The same techniques, MLP, GB, RF, AB, AB-LR, and AB-RF, are applied on the modified average dataset, and it was found that the boosting algorithm (AB) gives the best F-Measure/ F-Score value. The same classification techniques have again been utilized on the hourly dataset, and the result obtained was compared with that of the result of the average dataset. The pre-monsoon hourly dataset was best suited for thunderstorm day classification using the boosting algorithm.

Abstract Rotameters are reliable primary sensors for measuring the flow rate of fluids in pipelines. However, they have the limitation of being local, indicative sensors that rely on vertically mounted, breakable, transparent, or semi-transparent tapered tubes, suitable only for transparent liquids. To enable remote indication of flow rates or for control purposes, a non-contact secondary sensor is necessary. This paper investigates a modified non-contact rotameter transducer using a magneto-resistive (MR) element as the secondary sensor. A signal conditioner circuit is designed to transmit the rotameter reading and transducer output to a remote location for monitoring and control. The theoretical equations explaining the sensor, transducer, and transmitter’s performance are derived and discussed. Experimental studies on the prototype transmitter have been conducted, with results showing good linearity and repeatability, validating the theoretical model. Furthermore, artificial neural networks (ANN) have been applied to optimize the flow rate measurement by enhancing accuracy and reducing any environmental impacts on sensor performance. ANN-based models provide better calibration and fault tolerance, ensuring improved precision in flow rate detection even under varying environmental conditions. The experimental results and their correlation with the theoretical equations are presented, confirming the effectiveness of the proposed system.

We provide a new seventeen-level cross-connected T-type (CT-type) inverter with a lower device count (RDC) in this study. Just eight power switches, four self-balancing capacitors, and two DC sources are needed for the suggested design. It reduces the voltage stress on the switches by avoiding the need for an H-bridge to generate the negative half cycle, in contrast to traditional multilevel inverters. The distinctive feature of this topology is its asymmetrical cross‑connected structure, enabling seventeen voltage levels with far fewer components while maintaining natural capacitor voltage balancing. By cascading capacitors and DC sources with the switches, the inverter achieves the desired output levels efficiently. Thanks to its strong self‑balancing and high‑voltage handling capability, the design is particularly well-suited for renewable energy integration into the grid. The gate driver signals are generated using level‑shifted pulse width modulation. A comparative study demonstrates the advantages of the proposed topology over existing designs in terms of component count, driver requirements, and total standing voltage. Its practicality is confirmed through both simulation studies and a laboratory prototype.

Self-cleaning cementitious composites are one of the most functional building materials that have recently gained significant importance in order to enhance the aesthetic durability, environmental characteristics and service life of building and infrastructure. The materials are intended to prevent dust, organic impurities, dirt and airborne contaminants from settling on open surfaces by photocatalytic or superhydrophobic effect or by a combination of both effects. Recent developments have been based on the introduction of nanomaterials (titanium dioxide, silica-based modifier, and surface treatment) in cement-based matrices, which are not only mechanically strong but also active on the surface. Even with laboratory success, there are the challenges of durability, cost, dependence on weather conditions, compatibility with existing construction practices, as well as long-term maintenance performance that yet to be overcome in practice. The field is shifting from proof-of-concept research to optimization for real-world applications, notably in facades, pavements, tunnels and urban air-purifying structures. This paper summarizes the latest advances in SCCs, highlights the significant technical and practical challenges, and outlines potential future prospects for increased use. The development of materials, standard testing procedures, and life-cycle assessment will be essential to making these new composites a viable construction option, it's claimed.

This study developed an optimized neural-network framework for predicting the factor of safety (FOS) against soil liquefaction using shear wave velocity and relevant geotechnical and seismic parameters. Three models were evaluated: artificial neural network (ANN), ANN optimized with a genetic algorithm (ANN-GA), and particle swarm optimization (ANN-PSO). A dataset of 300 randomly generated cases was developed using seven input variables: stress-normalized shear wave velocity, soil deposit depth, groundwater table depth, fines content, bulk density, peak ground acceleration, and earthquake magnitude. Model performance was assessed using the coefficient of determination (R²), and root mean square error (RMSE). Among the investigated models, ANN-PSO exhibited the best predictive performance. During training, it achieved an R² = 0.998 and RMSE = 0.004, while the testing phase yielded an R² = 0.972 and RMSE = 0.010. Its predictive capability was further evaluated through rank analysis, regression analysis, reliability assessment, error-matrix evaluation, and Taylor diagrams. The proposed models were also compared with conventional liquefaction assessment indices, including the liquefaction potential index and liquefaction severity index. Overall, the results showed that ANN-PSO provides an accurate, efficient approach for estimating liquefaction-related FOS and has potential for rapid seismic risk assessment and practical geotechnical engineering applications.