The purpose of this research is to develop an efficient single-phase grid-connected PV system using a better performing asymmetrical multilevel inverter (AMI). Circuit component reduction, harmonic reduction, and grid integration are critical criteria for better inverter efficiency. The proposed inverter’s optimized topology requires seven unidirectional switches, three symmetric dc sources, and three diodes to produce an asymmetric fifteen level output; whereas, the same configuration will generate 7, 11, and 15-level output with an appropriate choice of dc source magnitudes. It is possible to reduce inverter losses and boost efficiency by decreasing the number of switches used. The integration of an asymmetric 15-level inverter with a grid-connected solar photovoltaic system is discussed in this article. A grid-connected solar photovoltaic (GCSPV) system is modelled and simulated using an asymmetric 15-level inverter. The dc sources of the 15-level inverter are replaced with PV sources. The results were analyzed with different operating temperatures and solar irradiance conditions. The GCSPV system is controlled by a closed-loop control system using Particle Swarm Optimization (PSO), Harris Hawk Optimization (HHO), and Hybrid Particle Swarm Optimization-Genetic Algorithm (PSOGA) based Proportional plus Integral (PI) controllers. Grid voltage, grid current, grid power, and total harmonic distortion (THD) of grid currents were analyzed. The performance of the 15-level asymmetric inverter was evaluated by comparing the THD of the grid current and the efficiency of the grid-connected photovoltaic system.

Deep learning models are frequently implemented as black boxes, which restricts their application in sensitive sectors such as agriculture, despite their ability to achieve high classification accuracies. To resolve this issue, we introduce a comprehensive interpretability framework that implements Grad-CAM, Grad-CAM++, and LIME techniques on the most effective CNN + ViT + SVM ensemble model for the detection of citrus diseases. The proposed model obtained classification accuracies of 99% and 97% by augmenting the real-world Lemon and Orange datasets with 1600 and 4000 images, respectively. Interpretability was implemented on the DenseNet121 (Orange) and InceptionV3 (Lemon) backbones. The model focused on specific textures related to diseases, which was shown by the detailed image analysis from LIME, while Grad-CAM and Grad-CAM ++ highlighted clear areas of lesions with high detail. A perfect score of 1.00 over all assessed courses was achieved through quantified assessment using Intersection over Union (IoU) and Dice Coefficient, and confirmed accuracy in visual explanation. In addition, intra-class compactness and inter-class separability were confirmed using feature-space t-SNE mappings. This visual explanation framework enhances transparency and builds user trust in machine learning-based disease management decision making in practice through clearly interpretable visual explanations for both multi-class and binary classification tasks. This research is a major advance towards developing trustworthy and comprehensible AI (XAI) solutions for agricultural diagnosis, through providing visual explanations to both multi-class and binary classification tasks. The approach is scalable and model-independent and is capable of integration into real-time and mobile decision support tools.

Nowadays in the era of electronic media it is very critical to protect and conceal information. The project makes it better how we conceal confidential information in pictures through a system called image steganography. The idea is to ensure that, the concealed information remains secure and yet the image portrays as being normal and intact to the human eye. We are employing smart algorithms, encryption and the use of technologies such as Python based OpenCV and NumPy to conceal the data in such manner that no intruder can quickly recognize. A web based platform is also a part of the system as the users can transfer images easily as well as embed data and download data as it is secure. The method can be used to achieve high security and decent visual quality, thus, it is applicable in the real world such as encrypted communication or encrypted file sharing. Keywords: Image Steganography, Visual Quality, Data Hiding, Encryption, LSB, Python, OpenCV, Web Application, Steganalysis Resistance, Secure Communication.

This paper aims to minimize total harmonic distortion (THD) in a reduced-switch multilevel inverter using an advanced hybrid optimization framework. A 15-level asymmetric multilevel inverter (MLI) with a binary-weighted (1:2:4) DC source configuration is considered the test platform because it balances power quality and hardware complexity with only seven active switches, three DC sources, and three diodes. While such reduced-switch topologies are not new in the field, their harmonic performance can be critically limited by the selection of the optimal switching angles. To address this issue, Selective Harmonic Elimination Pulse Width Modulation (SHEPWM) has been adopted, where a hybrid metaheuristic optimization algorithm solves the nonlinear transcendental equations- Enhanced Whale Optimization Algorithm (EWOA), Adaptive Particle Swarm Optimization-Newton Raphson (APSO-NR), Particle Swarm Optimization-Genetic Algorithm (PSO-GA) and Harris Hawks-Differential Evolution (HH-DE). The main contribution of this paper is the systematic comparison and implementation of these hybrid algorithms using a single SHEPWM scheme, analyzing their convergence behaviour, statistical performance, and harmonic minimization. From the simulation studies, the HHDDE algorithm achieved the fastest convergence rate and a THD of 5.33%, with current THDs below 3%, ensuring compliance with the IEEE-519 harmonic standards without external filtering.