Engineering Proceedings

Other

15 pages, 1454 KB

Open AccessProceeding Paper

Physics-Regularized Neural Networks for Photovoltaic Power Prediction Under Limited Experimental Data

by Aswin Karkadakattil

Eng. Proc. 2026, 138(1), 1; https://doi.org/10.3390/engproc2026138001 - 20 Apr 2026

Viewed by 787

Accurate photovoltaic (PV) power prediction under limited experimental data remains a significant challenge, particularly when purely data-driven models generate predictions that violate fundamental physical constraints. This study proposes a physics-regularized neural network framework for data-efficient PV power modeling using only 45 real experimental [...] Read more.

Accurate photovoltaic (PV) power prediction under limited experimental data remains a significant challenge, particularly when purely data-driven models generate predictions that violate fundamental physical constraints. This study proposes a physics-regularized neural network framework for data-efficient PV power modeling using only 45 real experimental measurements of irradiance and temperature. To address data sparsity while preserving physical realism, a physics-guided synthetic augmentation strategy is introduced to generate additional training samples strictly within experimentally validated operating bounds. The proposed Physics-Informed Neural Network (PINN) incorporates two complementary physical constraints directly into the training objective: (i) enforcement of the Shockley–Queisser thermodynamic efficiency limit to maintain compliance with theoretical conversion bounds and (ii) monotonicity regularization to ensure non-negative power gradients with respect to irradiance. Unlike conventional post-processing correction methods, these physical constraints are embedded during model training, enabling simultaneous improvement in predictive accuracy and physical consistency. When benchmarked against a structurally identical unconstrained Artificial Neural Network (ANN), the proposed framework achieves strong predictive performance (R² = 0.9947, RMSE = 5.21 W) while reducing monotonicity violations by approximately 82%. Robustness evaluations under extrapolated irradiance conditions and elevated temperature scenarios further demonstrate stable and physically admissible behavior beyond the training domain. Overall, the results demonstrate that integrating limited experimental measurements with embedded physical priors enables reliable and physically consistent PV power prediction in sparse-data environments, highlighting the potential of physics-regularized learning for renewable energy modeling applications. Full article

► Show Figures

Figure 1

8 pages, 3306 KB

Open AccessProceeding Paper

Automated Response Surface Methodology: Computational Replication and Validation Framework for Optimizing Supercapattery Materials

by Thiago Ferro de Oliveira and Simoni Margareti Plentz Meneghetti

Eng. Proc. 2026, 138(1), 2; https://doi.org/10.3390/engproc2026138002 - 20 Apr 2026

Viewed by 652

Abstract

Combining Response Surface Methodology (RSM) with Central Composite Design (CCD) is a powerful statistical approach to optimizing materials in energy storage systems. This study presents an open-source Python (v3.8+) framework that replicates and validates the RSM-based optimization of NiCo₂S₄–graphene [...] Read more.

Combining Response Surface Methodology (RSM) with Central Composite Design (CCD) is a powerful statistical approach to optimizing materials in energy storage systems. This study presents an open-source Python (v3.8+) framework that replicates and validates the RSM-based optimization of NiCo₂S₄–graphene supercapattery materials. We validated the framework by replicating a 20-experiment CCD analyzing graphene/NCS ratios, hydrothermal time, and S/Ni molar ratios. Advanced optimization using the Differential Evolution algorithm was integrated to efficiently solve the high-dimensional response surface space. The model explained 97.16% of the variance, and comprehensive diagnostic tests confirmed the assumptions of normality and residual independence. This approach provides an open-source methodology that supports reproducible and scalable data-driven material design and facilitates transparent computational materials science studies. Full article

► Show Figures

Figure 1

12 pages, 4066 KB

Open AccessProceeding Paper

Advancements in Artificial Intelligence for Renewable Energy Systems over the Past Decades

by Md. Nurjaman Ridoy and Sk. Tanjim Jaman Supto

Eng. Proc. 2026, 138(1), 3; https://doi.org/10.3390/engproc2026138003 - 24 Apr 2026

Viewed by 656

Abstract

Sunlight, air, and other natural resources are invaluable gifts that must be utilized responsibly to enhance human welfare while preserving the environment and protecting all forms of life. The reliance on fossil fuels has increasingly threatened these resources, which has made the exploration [...] Read more.

Sunlight, air, and other natural resources are invaluable gifts that must be utilized responsibly to enhance human welfare while preserving the environment and protecting all forms of life. The reliance on fossil fuels has increasingly threatened these resources, which has made the exploration of sunlight and wind energy as major renewable energy sources a critical focus of research and development. Artificial intelligence (AI), originally developed to mimic human thought and decision-making processes, has become a transformative force in renewable energy systems by optimizing energy generation, management, and distribution for greater efficiency and sustainability. This paper shows the evolution of AI applications in wind, solar, geothermal, hydro, bioenergy, and hybrid energy systems over the last few decades. A bibliometric analysis of the literature was conducted systematically by reviewing relevant journal articles between 2000 and 2025. The analysis identifies research trends, collaboration patterns, emerging domains, and future directions. Different studies show that AI technologies’ capabilities improve several aspects of renewable energy for the purpose of integrating operations into the grid for users, specifically forecasting, improving system stability and frequency, and enabling transient stability assessment. The study highlights key challenges and provides high-level insights to guide future research and support the continued development and application of AI in renewable energy systems. Full article

► Show Figures

Figure 1

15 pages, 873 KB

Open AccessProceeding Paper

AI-Enhanced Strategies for Energy-Efficient Urban Environments

by Sk. Tanjim Jaman Supto and Md. Nurjaman Ridoy

Eng. Proc. 2026, 138(1), 4; https://doi.org/10.3390/engproc2026138004 - 7 May 2026

Viewed by 747

Abstract

Artificial intelligence (AI) is rapidly redefining the management of urban energy systems by coupling predictive analytics with closed-loop control across buildings, power grids, and mobility networks, positioning cities as critical leverage points in global decarbonization efforts. Contemporary urban environments generate vast, heterogeneous datasets [...] Read more.

Artificial intelligence (AI) is rapidly redefining the management of urban energy systems by coupling predictive analytics with closed-loop control across buildings, power grids, and mobility networks, positioning cities as critical leverage points in global decarbonization efforts. Contemporary urban environments generate vast, heterogeneous datasets that enable advanced machine learning applications; however, limitations remain, including interpretability–fairness trade-offs, fragmented data governance, interoperability gaps, cybersecurity risks, and insufficient long-term validation across diverse climatic and socio-economic contexts. This review evaluates AI-driven strategies for energy-efficient urban systems and identifies the technical and governance conditions required for scalable impact. The evidence synthesized indicates that supervised and ensemble learning models achieve high predictive accuracy for electricity demand and chiller performance, with models such as Random Forest Regression achieving R² values up to 0.9835 in electricity consumption forecasting, while unsupervised approaches detect latent inefficiencies in HVAC systems, delivering measurable savings typically around 6% under controlled benchmarking conditions. Deep learning architectures improve multi-building forecasting and real-time control, with hybrid CNN–LSTM models achieving prediction accuracies up to 97% and outperforming traditional statistical approaches in weekly energy demand forecasting achieving higher prediction accuracy and significant energy savings in complex urban subsystems with reported reductions of approximately 21–23% in residential and educational buildings and up to 37% in office HVAC systems. Hybrid and physics-informed AI models embed thermodynamic principles into data-driven frameworks, improving robustness, interpretability, and generalization. IoT sensor networks and edge-computing architectures support adaptive HVAC, demand–response, and smart-grid management, while integrated building–grid–mobility systems enhance load balancing, storage use, and carbon reduction. AI-enhanced strategies offer a credible pathway toward measurable reductions in urban energy use and emissions with deep reinforcement learning in digital twin environments reducing HVAC energy demand by 10–35% while maintaining thermal comfort within ±0.5 °C in line with ASHRAE standards, and overall energy savings reaching up to 44% in optimized systems when supported by interoperable infrastructures, secure digital-twin architectures, and standardized measurement and verification protocols. Full article

► Show Figures

Figure 1

9 pages, 1122 KB

Open AccessProceeding Paper

Multi-Objective Evolutionary Prediction with an Artificial Intelligence-Based Approach for Urban Energy Planning

by Md Rakibul Islam, Aritra Islam Saswato and Md Salah Uddin

Eng. Proc. 2026, 138(1), 5; https://doi.org/10.3390/engproc2026138005 - 26 May 2026

Viewed by 306

Abstract

This study investigates the relationship between weather conditions (temperature, humidity), air pollutants (PM_2.5, PM₁₀, and CO), and photovoltaic (PV) degradation characteristics using location-specific machine learning frameworks. A data augmentation technique was employed to enhance the predictive modeling datasets. The [...] Read more.

This study investigates the relationship between weather conditions (temperature, humidity), air pollutants (PM_2.5, PM₁₀, and CO), and photovoltaic (PV) degradation characteristics using location-specific machine learning frameworks. A data augmentation technique was employed to enhance the predictive modeling datasets. The research evaluates four machine learning models: AdaBoost, Gradient Boosting, Decision Tree, and Random Forest. We found strong regression analysis values using the addressed machine learning models. Furthermore, feature importance analysis reveals that PM_2.5 has the most significant impact on PV module degradation. Full article

► Show Figures

Figure 1

10 pages, 2706 KB

Open AccessProceeding Paper

Modelling and MATLAB-Based Optimisation of Carbon Dioxide Adsorption Using Zn-MOF-5

by Shonisani Salvation Muthubi, Dorcas Museme Mabulay and Pascal Kilunji Mwenge

Eng. Proc. 2026, 138(1), 6; https://doi.org/10.3390/engproc2026138006 - 22 May 2026

Viewed by 531

Abstract

The growing concern over greenhouse gas emissions has prompted the need for efficient carbon dioxide (CO₂) capture technologies. This study focuses on simulating CO₂ adsorption using a zinc-based metal–organic framework (Zn-MOF-5). The primary aim is to develop and refine a [...] Read more.

The growing concern over greenhouse gas emissions has prompted the need for efficient carbon dioxide (CO₂) capture technologies. This study focuses on simulating CO₂ adsorption using a zinc-based metal–organic framework (Zn-MOF-5). The primary aim is to develop and refine a robust MATLAB-based approach for equilibrium and kinetic modelling using the Linear Driving Force (LDF) model and Langmuir isotherm, capable of accurately predicting CO₂ adsorption performance under varying operational conditions. By employing advanced computational methods, this research seeks to streamline the process design and enhance the feasibility of sustainable CO₂ capture solutions. Excel was used for statistical analysis and validation, while MATLAB R2025a was utilised for equilibrium and kinetic modelling using the LDF model and the Langmuir isotherm. The independent effects of temperature, pressure, and flow rate were evaluated using the variable effect method. The study found a significant negative association between temperature and CO₂ uptake, consistent with the exothermic nature of the adsorption process. Pressure had a significant impact on adsorption, whereas flow rate had little effect within the investigated range. The simulated CO₂ uptake (21.196 mmol/g) closely matched the experimental data (21.07 mmol/g) with a 0.59% variance, validating the model’s trustworthiness. The research shows that Zn-MOF-5 has a strong adsorption potential and that simulation tools can significantly minimise experimental costs and time. Furthermore, it underscores the potential of simulation tools to significantly reduce experimental costs and time, paving the way for more efficient and effective carbon capture solutions. This initiative not only contributes to optimising process design but also promotes sustainable practices in addressing global CO₂ emissions. By contributing to process optimisation, this study aligns with the United Nations Sustainable Development Goal (SDG) 13: Climate Action, which emphasises the urgent need for innovative solutions to combat climate change and its impacts. Furthermore, it promotes sustainable practices to address global CO₂ emissions, thereby supporting broader efforts for environmental sustainability. Full article

► Show Figures

Figure 1

15 pages, 2816 KB

Open AccessProceeding Paper

The Role of Artificial Intelligence in Driving Renewable Energy Transition: From the Current Landscape to Future Pathways

by Md. Nurjaman Ridoy, Sk. Tanjim Jaman Supto, Gaurob Saha and Sabbir Hossain

Eng. Proc. 2026, 138(1), 7; https://doi.org/10.3390/engproc2026138007 - 22 May 2026

Viewed by 1135

Abstract

The shift from fossil fuels to renewable energy is a key component in achieving global sustainability and dealing with climate change. Natural resources, such as sunlight, air, water, and biomass, have tremendous potential to create clean energy; however, exploiting these resources in an [...] Read more.

The shift from fossil fuels to renewable energy is a key component in achieving global sustainability and dealing with climate change. Natural resources, such as sunlight, air, water, and biomass, have tremendous potential to create clean energy; however, exploiting these resources in an efficient, stable, and large-scale integration manner is difficult due to their variable and distributed nature. Artificial intelligence (AI) approaches that mimic human learning and decision-making have recently become viable approaches to solving renewable energy problems. This study mainly examines the current landscape of AI applications across solar, wind, hydro, geothermal, ocean, hydrogen, bioenergy, and hybrid energy systems. AI enhances renewable energy forecasting, improves power system frequency analysis and stability assessments, and optimizes dispatch and distribution networks. Beyond technical optimization, AI also contributes to broader sustainability goals, including energy efficiency improvements, intelligent smart grid management, and enabling mechanisms such as carbon trading and circular economy practices to reduce exposure to climate extremes. Drawing on evidence from a range of renewable energy areas, this review highlights the importance of AI in bridging the link between technological innovation and sustainable energy management. This paper discusses potential future research avenues, such as building sophisticated AI designs, energy-efficient chips, and data communication networks. Ultimately, the synergy between AI and renewable energy systems represents not only a technological advancement but also a transformative pathway toward a resilient, low-carbon future. Full article

► Show Figures

Figure 1

10 pages, 1570 KB

Open AccessProceeding Paper

Circular Design as a Key Strategy to Cut Embodied Energy: A Digital AI Tool to Support Materials and Data Exchange for a Sustainable Built Environment

by Gabriele Rossini, Paola Altamura and Serena Baiani

Eng. Proc. 2026, 138(1), 8; https://doi.org/10.3390/engproc2026138008 - 29 May 2026

Viewed by 238

Abstract

The NPRR research project “From waste to manufacturing” developed an AI-powered digital tool to support the transition towards a circular built environment in Italy. The tool integrates a web platform enabling data exchange about materials with recycled content between designers, manufacturers and waste [...] Read more.

The NPRR research project “From waste to manufacturing” developed an AI-powered digital tool to support the transition towards a circular built environment in Italy. The tool integrates a web platform enabling data exchange about materials with recycled content between designers, manufacturers and waste recyclers, with a CAD plug-in for real-time sustainability assessment. As such, the tool fosters the use of recycled materials and allows a reduction in embodied energy. AI, trained through scenario-based learning and stakeholder participation, assists designers in sourcing recycled materials and processing data. With further training by LCA experts, it could interpret Environmental Product Declarations to guide material selection in line with international regulations. Full article

► Show Figures

Figure 1

15 pages, 4192 KB

Open AccessProceeding Paper

Adaptive Neuro-Fuzzy Control of a Small Wind Turbine–Battery DC Microgrid for Remote Electrification in Uzbekistan

by Botir Usmonov, Ulugbek Muinov, Komil Usmanov and Nigina Muinova

Eng. Proc. 2026, 138(1), 9; https://doi.org/10.3390/engproc2026138009 - 1 Jun 2026

Viewed by 269

Abstract

Rural regions of Uzbekistan experience continuing issues of energy access because of poor grid networks and variable renewable sources. The solution is small-scale wind turbines and energy storage. But the wind speeds and load demand are variable, and thus this solution needs intelligent [...] Read more.

Rural regions of Uzbekistan experience continuing issues of energy access because of poor grid networks and variable renewable sources. The solution is small-scale wind turbines and energy storage. But the wind speeds and load demand are variable, and thus this solution needs intelligent control systems to perform its best. This paper is an attempt to design an adaptive neuro-fuzzy inference system (ANFIS) controller to control a small wind power system with a battery storage unit. The controller will be intelligent to control the flow of power between the wind turbine, battery, and local loads. A model of MATLAB/Simulink is created to simulate the reaction of the system to various wind and load conditions. The simulation results indicate that the ANFIS controller improves voltage regulation, reduces power fluctuations, and enhances battery charge–discharge performance compared to the conventional PI controller. Environmental variability is effectively responded to by the system, making it more reliable and energy-efficient. ANFIS control and wind–battery microgrid integration provides a feasible and expandable off-grid electrification solution to remote areas. This strategy promotes the renewable energy ambitions of Uzbekistan and offers an example of smart microgrid implementation in other resource-limited rural areas. The next steps would be towards practical applications and hardware verification. Full article

► Show Figures

Figure 1

13 pages, 4239 KB

Open AccessProceeding Paper

Analysis of Power System Stability Indices Concerning High Penetration of Renewable Energies

by Amel Brik, Nour El Yakine Kouba and Ahmed Amine Ladjici

Eng. Proc. 2026, 138(1), 10; https://doi.org/10.3390/engproc2026138010 - 1 Jun 2026

Viewed by 521

Abstract

Currently, the large-scale integration of renewable energy sources (RESs), such as wind turbines and photovoltaic array, is profoundly altering the dynamic behavior of power systems. In particular, the reduction in system inertia makes transient stability more critical and increases the sensitivity of the [...] Read more.

Currently, the large-scale integration of renewable energy sources (RESs), such as wind turbines and photovoltaic array, is profoundly altering the dynamic behavior of power systems. In particular, the reduction in system inertia makes transient stability more critical and increases the sensitivity of the network to disturbances. The originality of this work lies in the systematic analysis of the nonlinear dynamics of power systems by thoroughly examining the impact of RESintegration on system stability, particularly through frequency response and voltage profile. In this context, a methodology for the evaluation and optimization of power system stability was proposed, based on two key indicators: the Critical Clearing Time (CCT) and the Rate of Change of Frequency (RoCoF). The IEEE 39-bus test system was used as a benchmark to simulate different scenarios. Three-phase faults are applied to determine the corresponding CCT values and to assess the system’s ability to regain a stable operating state after a severe disturbance. In addition, RoCoF variations are analyzed to quantify the impact of RES penetration on the frequency stability of the network. The obtained results show that a high penetration of renewable energy sources tends to reduce the CCT and increase the RoCoF, indicating a reduction in the dynamic robustness of the system. These observations are confirmed through comparative simulations performed with and without renewable energy integration. In conclusion, this study highlights the importance of optimal placement of renewable generation units, as well as the use of the CCT and RoCoF indices as effective diagnostic and optimization tools for modern power systems characterized by a high penetration of renewable energy sources. Full article

► Show Figures

Figure 1

10 pages, 3469 KB

Open AccessProceeding Paper

Development of a Framework for Using AI in Building Façade Optimization: An Application Focusing on Retrofitting NYCHA Midrise Housing in New York City

by Beatriz Bordignon Cypriano

Eng. Proc. 2026, 138(1), 11; https://doi.org/10.3390/engproc2026138011 - 3 Jun 2026

Viewed by 208

Abstract

Artificial intelligence (AI) technology and the Industrial Revolution (4IR) have the potential to rapidly advance smart buildings, materials, and construction processes to meet global decarbonization goals. Current retrofit techniques can vary a great deal in terms of their methodology from place to place, [...] Read more.

Artificial intelligence (AI) technology and the Industrial Revolution (4IR) have the potential to rapidly advance smart buildings, materials, and construction processes to meet global decarbonization goals. Current retrofit techniques can vary a great deal in terms of their methodology from place to place, but most have in common time constraints, budgets, and the need to reduce energy usage. The objective of this study is to develop a framework that optimizes the façade retrofitting process with the help of AI, bringing it in as a decision-making tool that also accounts for other parameters, but differently to traditional retrofit methodologies, the tenants are the judges. Thus, this research explores the possibilities of applying the AI tool Midjourney for problem solving in retrofitting in the design stages of the process. The overall framework is developed by addressing materiality, high performance analysis, affordable costs, and user experience inputs. The effectiveness of the framework is tested for three selected façade system options in terms of their Energy Use Intensity (EUI) simulated on Grasshopper, showcasing a final EUI reduction for options A, B and C analyzed at 32%, 5.8% and 26%, respectively. In addition, Life Cycle Analysis (LCA) and the overall costs were also simulated and compared across options. Full article

► Show Figures

Figure 1

8 pages, 1663 KB

Open AccessProceeding Paper

From Solar Panels to AI Decisions: Intelligent Server Utilization for Sustainable Computing

by Nikolaos Fragkos, Stylianos Katsoulis, Evangelos Nannos, Fotios Zantalis, Ioannis Chrysovalantis Panagou, Panagiotis Tsiakas and Grigorios Koulouras

Eng. Proc. 2026, 138(1), 12; https://doi.org/10.3390/engproc2026138012 - 25 Jun 2026

Viewed by 160

Abstract

Renewable integration is increasingly important for sustainable off-grid computing. The inherent variability of solar output frequently produces unusable midday surpluses. Leveraging recent Artificial Intelligence (AI) advances and established literature, we evaluate an AI-driven demand-response framework for scaling Large Language Models (LLMs) training servers [...] Read more.

Renewable integration is increasingly important for sustainable off-grid computing. The inherent variability of solar output frequently produces unusable midday surpluses. Leveraging recent Artificial Intelligence (AI) advances and established literature, we evaluate an AI-driven demand-response framework for scaling Large Language Models (LLMs) training servers using real-time solar energy data, Solcast forecasts, and battery storage records collected from Battery Management Systems (BMS), Maximum Power Point Tracking (MPPT) units, and smart inverters. An n8n AI Agent using the Ollama chat model gpt-oss:20b assesses surplus solar energy, activating selected servers to utilize otherwise wasted capacity. Workloads consistently align with solar availability, demonstrating 99% operational reliability, sub-second responsiveness, and accurate surplus-energy detection. This research demonstrates how Artificial Intelligence can repurpose surplus solar output into usable computational capacity, thereby contributing to a broader transition toward renewable-powered infrastructures. Full article

► Show Figures

Figure 1

Journal Menu

Journal Browser

Eng. Proc., 2026, Designs 2026

Other

Further Information

Guidelines

MDPI Initiatives

Follow MDPI