Al Akhawayn University

The growing global energy consumption, driven by factors such as population growth, industrial expansion, and the increasing electrification of sectors like transportation, makes it crucial to have robust models that predict electricity usage accurately. This rise in energy demand has made the power systems unable to balance energy supply and demand. Hence, we need accurate electricity forecasting to ensure that the right amount of energy is generated and distributed to prevent power outages, minimize energy waste, and optimize resource use. This is particularly important in modern energy grids, where the integration of renewable energy sources like solar and wind introduces variability and unpredictability. Nonetheless, electricity load forecasting has been traditionally reliant on statistical models like autoregressive integrated moving average (ARIMA) and seasonal autoregressive integrated moving average (SARIMA) which recently have started becoming obsolete with the demand structure of energy which has changed drastically in the recent past. The emergence of machine learning (ML) and deep learning (DL) methods saved the situation nearly completely. Models like Artificial Neural Networks (ANNs), Support Vector Machines (SVMs), and Long Short-Term Memory (LSTM) networks among others are able to understand and work with such complex and non-linear data. However, challenges persist, including the need for real-time data, enhancing data quality, and addressing the "black box" nature of some models, which can limit interpretability. This paper provides a comprehensive review of the latest methodologies for electricity consumption prediction, covering traditional statistical, machine learning, and deep learning approaches. It also examines recent trends, such as the integration of Internet of Things (IoT) devices for real-time data collection and the use of hybrid forecasting models. The review highlights key areas for future research, aiming to address the ongoing challenges and improve the accuracy and reliability of electricity consumption predictions.

Wheat productivity is increasingly threatened by numerous foliar diseases and pest infestations, making timely and accurate diagnosis essential for sustainable crop management. Although deep learning-based computer vision approaches have shown promise in plant disease recognition, most existing studies focus on a limited set of disease types or rely on datasets that do not capture real field diversity. This paper presents a comprehensive multi-class image classification study utilizing the Wheat Plant Diseases dataset, which comprises fifteen distinct categories covering major wheat diseases, insect pests, and healthy plants. Leveraging MobileNetV2 and VGG16 as baseline convolutional architectures, we investigate standardized training pipelines and extensive data augmentation protocols tailored to the complexities of wheat stress classification in realistic field conditions. Experimental results demonstrate high efficacy for both architectures: VGG16 achieved a classification accuracy of 96.28%, while the lightweight MobileNetV2 reached a competitive 93%. Notably, training dynamics revealed little to no overfitting, validating the robustness of the proposed data augmentation strategy. The study further provides a comparative analysis of model efficiency suitable for resource-constrained environments. This work establishes new baselines for comprehensive wheat disease and pest classification and provides actionable guidance for practical computer vision systems in precision agriculture.

The global shift toward sustainable transport electric vehicles (EVs) is at the core of decarbonization efforts. While advanced economies have achieved their rapid adoption through strong policies and incentives, emerging markets face structural and behavioral barriers. This study investigates the paradox in Morocco, whereby a significant automotive capacity contrasts with a minimal domestic BEV market share of 0.6%, despite 143% growth from a small base, using a four-dimensional framework encompassing financial, infrastructural and energy, policy and institutional, and behavioral–social factors. The research integrates a literature review, a survey (n = 522), and secondary data on charging infrastructure and EV sales. Findings reveal a strong value–action gap: 69% of respondents acknowledged EVs’ environmental benefits yet only 1.1% owned one and 42% had considered buying. The high upfront costs of EVs influenced over 70% of participants, and a significant association was confirmed between charging availability and purchase intent (χ2 = 34.80, p < 0.05). Urban-centric charging, fragmented governance, and skepticism persist as barriers. The study concludes that industrial strength alone cannot ensure adoption without targeted incentives, equitable infrastructure, and cultural shifts in ownership perception, offering key insights for policymakers in emerging economies pursuing sustainable mobility.

Expanded Polystyrene Aggregate Concrete (EPAC) offers reduced structural density and carbon footprint while maintaining acceptable mechanical performance, making it a promising material for sustainable, seismic-resistant construction. This study investigates the natural self-healing behavior of EPAC, a largely unexplored area, using the Taguchi method to analyze the effects of EPS percentage of total aggregate volume (60%, 70%, 80%), cement content (410, 515, 594 kg/m 3 ), and W/C (0.45, 0.5, 0.55). A novel method developed in this study, quantified a maximum healing efficiency of 67.5%, with results indicating that higher EPS content enhances healing due to its elastic nature. Compressive strength, thermal conductivity, and density were also assessed to validate the method’s reliability. The findings demonstrate the utility of the Taguchi method in construction materials research by reducing experimental workload while maintaining analytical depth. The proposed healing assessment method opens new avenues for evaluating durability in EPS-based concretes, supporting future innovations in sustainable construction.