Journal of Engineering in Industrial Research

Abstract Persistent efficiency limitations and manufacturing challenges in silicon photovoltaics necessitate innovative strategies. This study introduces an artificial intelligence AI-driven optimization framework for n-ZnO/TiO₂/p-Si solar cells using semiconductor-insulator-semiconductor (SIS) heterojunction engineering. A hybrid deep learning model DeepSeek combined with Bayesian-optimized SCAPS-1D simulations, tuned critical parameters: ZnO thickness, TiO₂/Si interface defects, and p-Si doping. The AI-optimized design achieved 16.93% power conversion efficiency (PCE) under AM1.5G illumination, closely aligning with the predicted 17.99%. Key innovations include an 80 nm ZnO layer to minimize resistive losses and enhance carrier extraction, paired with ultra-low TiO₂/Si interface defect density (5×10¹¹ cm⁻²). The SIS architecture employs TiO₂ as an electron transport layer and ZnO as a hole-blocking layer. AI analysis revealed a Type-II band alignment at the TiO₂/ZnO interface, synergistically enhancing open-circuit voltage (Voc) and fill factor (FF >79%). This defect-aware design suppresses carrier recombination, outperforming conventional PN-junction cells through superior carrier selectivity. The AI-driven workflow reduced computational costs by 15%, offering a scalable pathway for high-efficiency photovoltaics. By bridging theoretical modelling with experimental feasibility, this work highlights AI’s transformative potential in accelerating material discovery and device optimization. These advancements position AI as a cornerstone for next-generation, cost-effective solar technologies, accelerating the global transition to sustainable energy solutions.

Abstract The increasing global demand for energy, the depletion of fossil fuel reserves, and the environmental consequences of their consumption have made the need for transformation in energy supply systems more urgent than ever. In this context, renewable energy sources have emerged as sustainable alternatives, yet each faces challenges such as dependence on climatic conditions, high infrastructure costs, and storage limitations. Among these options, hydrogen has garnered significant attention from researchers and policymakers due to its high energy density, zero harmful emissions, storage capability, and diverse production sources. This study examines the energy crisis, the limitations of renewable energy sources, and the unique properties of hydrogen, analyzing its significance as a modern energy carrier. Various hydrogen production methods, including gray, blue, and green hydrogen, explored, along with the challenges related to production, storage, and distribution. In addition, the role of hydrogen in industries, transportation, electricity generation, and renewable energy storage was evaluated. The results indicate that with advancements in production technologies and cost reductions, hydrogen could become a key component of a sustainable energy future. Achieving this goal requires investment in infrastructure development, improvements in storage and transportation technologies, and the implementation of effective supportive policies. If existing challenges addressed, hydrogen could play a crucial role in reducing greenhouse gas emissions, enhancing energy security, and facilitating the transition to a sustainable, low-carbon energy system.

Abstract This study presents a statistical evaluation of GRMobiNet, a novel lightweight deep neural network model designed for efficient image classification in resource-constrained environments (RCEs). Built upon the MobileNet architecture, GRMobiNet introduces targeted modifications to enhance predictive accuracy without increasing model size and computational complexity. To validate the performance gains, a series of controlled experiments were conducted using a custom dataset comprising ten image categories, with each model evaluated across ten repeated runs on identical test sets. A paired samples t-test was applied to compare the classification accuracy of GRMobiNet and MobileNetV2 under identical experimental conditions. Results indicate that GRMobiNet achieved a mean accuracy of 80%, significantly outperforming MobileNetV2’s 57%, with a mean improvement of 2.3 correctly predicted images per run. The observed p-value of 0.019 confirms statistical significance at the 95% confidence level. Moreover, GRMobiNet exhibited lower variance and a reduced standard error of the mean, indicating greater stability across trials. These findings confirm that GRMobiNet offers not only computational efficiency, but also statistically validated performance superiority, making it highly suitable for real-world deployment in domains such as mobile diagnostics, precision agriculture, and embedded surveillance systems. The statistical rigor of this validation underscores GRMobiNet’s robustness and reliability as an advancement over existing lightweight architectures.

Abstract This study introduces a new technique that utilizes nanohybrids made from graphene ferrite (CoFe₂O₄) quantum dots nanohybrids (G-FQDs). These nanohybrids are produced through a manufacturing process that is both inexpensive and scalable for commercial use. When these FQDs-G nanohybrids are dispersed in a fluid to form nanofluids, they significantly lower the interfacial tension the force at the boundary between oil and water. Reducing this tension helps to improve the displacement of oil in reservoirs, making these nanofluids highly effective for enhanced oil recovery (EOR) technologies. The graphene synthesis involves chemical exfoliation using deionized water, strong acids, and oxidants, while ferrite quantum dots (FQDs) are produced via a hydrothermal method using castor oil plant precursors. The G-FQDs nanohybrids are fabricated using a sol-gel process and characterized using techniques such as XRD, FTIR, and HRTEM. We investigated the mechanism of these nanofluids under reservoir-simulated conditions. The results demonstrate that 0.5-G-FQDs nanofluid optimally modifies wettability in oil-wet carbonate slabs while exhibiting the optimal stability through minimal droplet formation. These nanohybrids significantly reduce interfacial tension, with oil/water IFT dropping from 14.5 mN/m to 1.80 mN/m and n-decane/water IFT, which is a standard hydrocarbon phase that mimics the properties of crude oil decreasing from 46.6 mN/m to 24.3 mN/m. This study confirms the potential of G-FQDs nanohybrids for enhancing EOR efficiency.

Abstract Gallium Arsenide Bismide (GaAsBi) is a III-V semiconductor alloy formed by adding bismuth (Bi) to gallium arsenide (GaAs), resulting in a reduced bandgap and improved optoelectronic properties. Its unique characteristics make it a promising material for advanced photodetector applications and high-efficiency solar cells. The structural, electronic, and optical properties of pristine Gallium Arsenide (GaAs) and its alloy GaAs_1-xBi_x (x = 0.25, 0.5, 0.75, 1) compounds were explored using the first principle approach with a full potential linear augmented plane wave (FPLAPW) method as implemented in WIEN2K code. The structural properties, including lattice constant, volume, and bulk modulus, were assessed after optimization using Murnaghan's equation of state. Electronic properties were determined using two methods such as the generalized gradient approximation in Perdew-Burke-Ernzerhof (PBE) and the modified Becke-Johnson (mBJ) approach for the exchange-correlation potential. Optical parameters including absorption coefficient, reflectivity, conductivity, and others were simulated from a computed dielectric function using the Kramers-Kronig relation.

Abstract Life cycle economic assessment serves a vital role in evaluating sustainability across various industries. This study aims to conduct a comparative life cycle economic analysis of different conversion technologies to identify the most viable options for treating Municipal Solid Waste (MSW). It specifically examines six different scenarios based on these conversion technologies: anaerobic digestion for heat and electricity, anaerobic digestion for household biogas, anaerobic digestion for fuel gas, pyrolysis for electricity, fermentation for bioethanol, and landfill gas collection for household use. A case study was conducted in Aradkooh, Iran. The results reveal that the landfill gas collection scenario is the most economically viable option, with a net present value (NPV) of $5.28 million. In contrast, the fermentation scenario shows an NPV of -$1,209 million, highlighting its significant economic shortcomings. The sensitivity analysis demonstrates that the most influential factors in enhancing the NPV are the operational efficiency of the conversion technologies and the market prices for bio-energies. It is highly recommended that stakeholders secure sufficient funding to improve treatment technology readiness. Additionally, the government should take proactive steps to attract venture capital in this sector by developing strong infrastructure and offering compelling incentives. To promote sustainable decision-making, it is essential to conduct comprehensive environmental and social assessments along with economic evaluations before implementing any conversion technologies.