Kathiravan M

Thermal systems are fundamental to a wide range of industrial applications, where performance and efficiency critically depend on precise and reliable modeling techniques. Traditional Artificial Neural Network (ANN)-based models, although widely used, often struggle with overfitting, limited generalization, and inadequate representation of the complex, nonlinear behavior inherent to thermal processes. These limitations restrict their deployment in real-time and dynamic operational settings. This study aims to enhance the predictive accuracy and robustness of thermal system modeling by integrating advanced machine learning (ML) techniques with hybrid optimization strategies. The research focuses on complex systems such as heat exchangers, gas-solid fluidized beds, and thermal energy storage units. A comprehensive methodology involving industrial data collection, preprocessing via normalization and feature selection, and model training using individual and hybrid ML algorithms is proposed. Performance is benchmarked using Mean Absolute Error (MAE), Root Mean Square Error (RMSE), and R² metrics. Advanced methods like Deep Learning (DL), Support Vector Machines (SVM), Genetic Algorithms (GA), Ensemble Learning, Transfer Learning, and Evolutionary Optimization are employed to address shortcomings of conventional approaches. Results demonstrate that hybrid models outperform standalone ANN-based techniques in prediction accuracy and generalization.

Inhaling formaldehyde a chemical that is widely used in many different industries can have serious health consequences. In order to precisely detect formaldehyde levels in industrial air quality environments, this study makes use of deep learning techniques. Using sensor data gathered from high-risk industrial areas the study focuses on variables like air quality index temperature and humidity. The data is processed by Convolutional Neural Networks (CNNs), which identify trends linked to increases in formaldehyde concentrations. To improve model accuracy preprocessing of the data is done including feature scaling and outlier elimination. The model's performance is assessed using evaluation metrics like Mean Squared Error (MSE), sensitivity, specificity, and prediction accuracy. Results show that when compared to conventional regression models the CNN-based model considerably lowers false positives while achieving a high prediction accuracy. Rapid reaction to hazardous formaldehyde levels is made possible by the deep learning frameworks' real-time monitoring capability which lowers possible health hazards. To improve long-term prediction accuracy and trend identification future research will investigate the use of recurrent neural networks (RNN) for time-series analysis.

Potholes pose a major risk to roadway safety and vehicle durability, necessitating timely detection and repair. With advancements in artificial intelligence, deep learning has become crucial for automating pothole detection and segmentation. Previous studies using CNN and Haar cascade methods have achieved accuracies up to 98.2%. This paper presents a novel approach leveraging the YOLOv8 architecture for instance segmentation, enhancing detection accuracy by capturing contextual and spatial relationships. The process involves data collection, annotation, preprocessing, and model training using datasets from Roboflow Universe and Kaggle. The model's performance is assessed through sensitivity, precision, recall, F1 score, and mean average precision (mAP). Experimental results indicate a significant improvement, achieving a 99.2% mAP in pothole detection and segmentation. These findings highlight the potential of YOLOv8 in advancing automated road maintenance, ensuring safer and more efficient transportation systems.

Text extraction from images plays a crucial role in optical character recognition applications such as invoices and receipt recognition. The recent character recognition approaches work well for good-quality scanned receipts, but they fail to do the same for low-quality receipts, offering reduced accuracy instead. This paper proposes invoice receipt identification using normalized cross-correlation-based template matching and a novel auto-text extraction approach using a deep learning algorithm. The proposed technique includes three major steps, preprocessing, character recognition and post-processing. The first step, which commences with preprocessing, involves noise removal, quality enhancement and image de-skewing. In the second step, auto-text extraction is carried out using a deep learning algorithm. The final post-processing step includes configuring the extracted text and exporting it to Word/Excel. According to the experimental results, the accuracy of the proposed approach outperformed existing approaches.

Introduction: WTI crude oil prices are influenced by multiple macroeconomic factors, primarily related to supply. This study employs regression equations and Monte Carlo simulations to analyze these dependencies and predict price trends. Objectives: This research aims to establish the relationship between WTI crude oil prices and eight key macroeconomic indicators. Historical data from the past 15 months is compiled to support this analysis. Methods: A multiple regression model is used to quantify the impact of macroeconomic variables. Multicollinearity is assessed to ensure model robustness, and Monte Carlo simulations are applied for scenario generation. Results: Findings reveal that certain macroeconomic variables significantly affect crude oil price fluctuations. The model successfully captures dependencies and provides insights into potential price movements. Conclusions: This study enhances understanding of crude oil price dynamics, aiding investors and policymakers in decision-making. The proposed framework can be extended for further analysis of commodity markets.