Sustainable Energy and Artificial Intelligence

Abstract Climate change has significantly reshaped energy supply and demand by increasing the frequency and severity of extreme events, thereby intensifying challenges related to the sustainability and security of energy systems. In response, artificial intelligence has emerged as a critical enabling technology with the potential to improve forecasting accuracy, operational flexibility, and system resilience under non-stationary climatic conditions. Nevertheless, the future integration of AI into energy systems is subject to profound uncertainties arising from climatic dynamics, governance structures, and policy environments. This study investigates plausible futures of AI-enabled climate-responsive energy systems in Iran through a scenario-based approach. The research adopts an applied mixed-methods design. Initially, key drivers influencing the future development of climate-responsive energy systems were identified through a systematic literature review and expert interviews. Subsequently, Cross-Impact Analysis using the MICMAC method was employed to examine the influence–dependence relationships among these drivers. The analysis identified climate and energy security pressures and environmental and market-based policy stringency as the dominant exogenous drivers shaping system evolution. Based on the alternative states of these two critical uncertainties, four plausible future scenarios were constructed. The results indicate that under high climate pressure, strong and coherent governance combined with market-oriented environmental policies can convert stress into a driver of structural transformation and systematic AI deployment. Conversely, weak policy frameworks even when climate pressures are moderate lead to missed opportunities and the accumulation of long-term vulnerabilities. Overall, the study highlights that managing uncertainty through effective governance and strategic AI adoption is central to achieving resilient energy futures.

Abstract Amid escalating environmental concerns and the imperative for sustainable resource management, optimizing recycling processes has become a paramount challenge. This research proposes an advanced methodology for sorting recyclable waste by integrating a robotic arm with four degrees of freedom and a conveyor system, guided by a genetic algorithm (GA). The system processes waste represented as color-coded cubes with assigned values, categorizing them based on predefined metrics to maximize efficiency.
To address the constraints of the robotic arm's limited operational range and the dynamic nature of the conveyor belt, a genetic algorithm with variable-length chromosomes was employed. This approach optimizes the sorting process by prioritizing high-value items while adhering to stringent temporal and spatial constraints. The methodology was simulated and validated using RoboDK software, with Python utilized for algorithm implementation.
The findings demonstrate substantial improvements in sorting efficiency and cumulative value compared to traditional sequential methods. This study underscores the potential of integrating robotic systems with intelligent optimization algorithms to advance industrial recycling operations, enhancing automation efficiency and sustainable recycling practices at an industrial scale.

Abstract Over the past decade, the development of advanced heating, cooling, and thermal energy storage systems has increasingly focused on innovative solutions such as nanofluids, metal foams and phase change materials (PCMs). Through the coordinated use of complementary experimental and modeling approaches. This project systematically assessed the degree of heat transfer and overall thermal performance enhancement achievable in heat exchangers and thermal energy storage systems by integrating nanofluids, metal foams, and PCMs. This article provides an overview of the main aspects of this comprehensive research activity. Particular attention is devoted to the comparison between mesoscopic and macroscopic heat transfer modeling in metal foams and nanofluids, as well as to the experimental datasets acquired and analyzed throughout the research program. Overall, this multiscale review indicates that the combined use of metal foams, nanofluids, and phase change materials can effectively enhance heat transfer characteristics and thermal energy storage performance, while also highlighting the remaining modeling and practical challenges associated with their implementation in engineering systems.

Abstract Modern Multi Terminal HVDC (MT HVDC) grids are expanding rapidly with large scale renewable integration, yet disturbance detection still faces challenges in achieving the required speed, adaptability, and selectivity. Recent advances highlight a shift toward data driven approaches that learn disturbance signatures directly from measurements, reducing reliance on handcrafted features, threshold tuning, and complex signal processing pipelines.
This study proposes a hybrid AI based disturbance detection method that integrates a lightweight LSTM network with an XGBoost classifier for real time operation in MT HVDC systems. The method relies solely on voltage measurements sampled at 1 kHz, significantly reducing data rate requirements while preserving the temporal resolution necessary for fast decision making. Voltage waveforms are segmented into 1 ms windows, from which integrated voltage and rate of change features are extracted and encoded into compact temporal embeddings by the LSTM. These embeddings are subsequently classified by XGBoost to provide fast, interpretable, and probabilistic disturbance decisions.
A detailed MT‑HVDC test system was developed in PSCAD to generate a wide range of fault scenarios and operating conditions. Comprehensive benchmarking and multi‑objective hyperparameter optimization identified an efficient single‑layer LSTM with hidden size 8 and batch size 32 as the optimal balance between accuracy and computational efficiency. Results demonstrate strong generalization capability and robust performance under non‑ideal conditions.
Overall, the proposed method offers a scalable, data centric and deployment ready solution that enhances disturbance detection speed, improves fault clearance verification, and contributes to reducing Total Fault Clearance Time in next generation HVDC protection architectures.

Abstract This paper addresses the problem of model-free control and sensor fault detection for discrete-time nonlinear systems by proposing an intelligent proportional–integral–derivative controller in conjunction with a fractional-order sliding mode observer. The proposed method eliminates the need for an explicit mathematical model and relies solely on input and output signals. By integrating sliding mode observation with fractional calculus, the proposed fault detection scheme benefits from enhanced memory characteristics and increased degrees of freedom in the observer design. This integration enables an effective trade-off between detection accuracy and convergence speed, thereby improving the overall performance of the fault detection mechanism. Enhancing the fault detection mechanism significantly reduces false alarm rates, thereby improving operational reliability and yielding tangible benefits in terms of economic efficiency. To intelligently tune the controller and observer parameters, a Reinforcement Learning–based optimization strategy is employed. This learning mechanism enables the adaptive tuning of controller and observer design parameters to achieve enhanced tracking accuracy while simultaneously improving the precision of residual generation for reliable fault detection within the system. The stability of the resulting closed-loop system is rigorously established using Lyapunov theory. The effectiveness and superiority of the proposed approach are validated through comprehensive simulation studies conducted on a data-driven, model-free nonlinear discrete-time system. The simulation results demonstrate significant improvements in tracking performance, robustness, and fault estimation accuracy compared to conventional approaches.

Abstract Natural disasters, such as floods and earthquakes, frequently cause widespread power outages and irreversible damage to equipment and consumers within the power industry. Ensuring a safe and uninterrupted electricity supply during such events represents a primary challenge for modern power system operators, a concept often termed network resilience. This paper proposes a network reconfiguration plan, integrated with photovoltaic (PV) resource management, to enhance the resilience of a power distribution network against natural disaster threats. The proposed approach minimizes the total expected costs, comprising both equipment repair/reinforcement and consumer outage costs, which serve as the objective function. Furthermore, the model incorporates AC power flow constraints, network reconfiguration logic, and PV resource capacity limits. Uncertainties—including active and reactive load demands, PV generation, and the availability of network components (such as main/reserve lines and PV resources)—are modeled using scenario-based stochastic programming. Scenarios are generated via the Monte Carlo method, with a subset selected using the Kantorovich reduction technique. Finally, the resulting stochastic optimization problem is formulated as a Mixed-Integer Nonlinear Programming model and solved using GAMS software. The proposed method is implemented and analyzed on 33-bus and 119-bus sample networks through four distinct case studies. Numerical results conclusively demonstrate that the integration of robust PV resources, combined with network reconfiguration, significantly improves network resilience under various natural disaster scenarios.