The Electrical Engineering Department is inviting undergraduate students to apply for research courses for Term 261:
For any questions, please contact the undergraduate research coordinator:
Dr. Abdulaziz Al-Qwbaiban - qwbaiban@kfupm.edu.sa
Project Manager: Dr. Amjad Ali amjad.ali@kfupm.edu.sa
Project Title: A Smart Agentic AI-Based Cleaning Scheduling Framework by Intelligent Soiling Prediction for Desert PV Systems
Project Field: Renewable Energy, Solar PV, Electrical Engineering, and AI
Project Description: The rapid growth of solar photovoltaic (PV) installations in Saudi Arabia has created a big operational problem: the dry conditions cause dust, sandstorms, and humidity to stick to the panels, which costs the company 0.3–1% of its energy every day and more than 40–60% during extreme events. Cleaning remains one of the most expensive and water-intensive maintenance tasks. However, present technologies rely on predefined schedules or unreliable assessments, which leads to too much water use, delays in action during heavy soiling, and wasted energy that could be avoided. The proposed project aims to address this gap by developing a comprehensive AI-driven cleaning management system that integrates forecasting, multi-sensor monitoring, aerial visual assessment, and optimisation through reinforcement learning. Data related to environmental, operational, and visual parameters from laboratory setups and a pilot photovoltaic array including irradiance, temperature, humidity, particulate concentrations, wind patterns, thermal images, and soiling metrics will be pre-processed and used to train deep learning models like Long Short-Term Memory networks, CNN-LSTM hybrids, and gradient-boosted etc models for short-term predictions of soiling behaviour and energy loss. A real-time fusion layer that uses Kalman filtering, weighted feature integration, and attention methods will combine thermal, optical, and image-derived characteristics to make a reliable soiling index. To make soiling maps, find hotspots, and measure uneven dust distribution across large photovoltaic fields, aerial images will be processed using segmentation and detection techniques like U-Net and YOLOv8. A reinforcement learning decision engine will determine the most cost-effective cleaning schedule by evaluating expected soiling, potential energy recovery, predicted weather, water availability, and cleaning costs. This will make it feasible to take actions like immediate cleaning, delayed cleaning, intervention in specific areas, or more inspection. All modules will be combined into a prototype that works with the Internet of Things (IoT). The prototype will then be further tested in a lab and in the field to see how well it saves water, increases energy yield, optimises cleaning frequency, and works reliably in real desert conditions. The expected results include a precise soiling prediction, a combined real-time soiling measure, an automated cleaning scheduler, advanced inspection tools, and a prototype system that works perfectly. This project will help Saudi Arabia reach its Vision 2030 goals by making renewable energy more efficient, using less water, and making it easier to maintain photovoltaic equipment using smart, data-driven methods. It also helps with scientific knowledge, intellectual property, and high-impact publications.
Project Manager: Dr. Naveed Iqbal naveediqbal@KFUPM.edu.sa
Project Title: Cardio-Reason: Multimodal Foundation Model for Language-guided Cardiac Segmentation and Reasoning-based Pathology
Project Field: Image Processing
Project Description: The predominant approach in automated cardiac image analysis relies on fully supervised convolutional neural networks (CNNs), particularly U-Net and its variants, which have demonstrated superior performance in standard segmentation tasks. Despite these advancements, such architectures exhibit significant limitations: they function as "black boxes" with limited interpretability, are confined to closed-set anatomical taxonomies established during training, and address segmentation and diagnosis as separate tasks. As a result, these models do not generalize effectively to rare pathologies or enable zero-shot segmentation of previously unseen anomalies without catastrophic forgetting. With the advancement of large language models (LLMs), it is essential to move beyond pixel-level classification and develop systems capable of high-level semantic reasoning. In this work, we propose Cardio-Reasoning, a novel multimodal large language model (MLLM) framework that unifies high-fidelity spatiotemporal segmentation with automated clinical reasoning. Unlike traditional methods that rely on 3D processing, Cardio-Reasoning leverages the capabilities of the foundation model, the segment anything model (SAM), as a visual backbone. By treating cardiac magnetic resonance imaging (MRI) as a 4D spatiotemporal sequence, the model preserves temporal coherence across the cardiac cycle, enabling precise quantification of dynamic biomarkers, such as strain and wall motion abnormalities, which are frequently lost in slice-by-slice processing. To bridge the domain gap between vision and cardiac physiology, we propose a parameter-efficient fine-tuning (PEFT) strategy utilizing low-rank adaptation (LoRA). This allows the foundation model to adapt to the specific textural and contrast properties of medical imaging (e.g., late gadolinium enhancement) while retaining its broad generalization capabilities, drastically reducing the computational burden compared to full-model training. The Cardio-Reasoning uses a language-instructed segmentation model that distinguishes between the embedding space of a clinical LLM and the visual backbone's mask decoder. Furthermore, the token strategy will be used to enable reasoning segmentation, which accepts complex, open-vocabulary clinical prompts (e.g., localize regions of subendocardial ischemia and correlate with the ejection fraction), and generates segmentation masks that are semantically aligned with the diagnostic text. This allows the LLM to utilize the geometric outputs of the vision encoder as visual evidence for chain-of-thought (CoT) reasoning, thereby grounding the generated pathology reports in verifiable anatomical data. The proposed research will demonstrate superior zero-shot performance on rare congenital defects and pathologies absent from standard training datasets. By integrating geometric precision with semantic reasoning, this research establishes a novel chat-to-diagnosis framework. This approach advances cardiac AI from static, task-specific tools to explainable, queryable agents, ensuring both diagnostic rigor and clinical transparency.
Project Manager: Dr. Omar Alkhazragi omar.khazragi@kfupm.edu.sa
Project Title: Monolithic AlGaAs Photonic Integrated Circuits with VCSELs for Quantum Technologies
Project Field: Optoelectronics, photonics, photonic integrated circuits
Project Description: This project is focused on a novel design for photonic integrated circuits for quantum applications. It integrates the laser and the passive photonic circuit components on the same chip monolithically. The project starts with the design step and simulation, then it will focus on testing the proposed design experimentally.
Project Manager: Dr. Ahmed Abdelmottaleb Abdelfattah Attia Omar ahmed.omar.1@kfupm.edu.sa
Project Title: AI-Assisted Forward Design for Optically Transparent Electromagnetic Wave Absorbers
Project Field: Applied electromagnetics
Project Description: Design a periodic surface with some electromagnetic response with the help of AI
Project Manager: Dr. Fahad Saleh Al-Ismail fsalismail@kfupm.edu.sa
Project Title: Design and Validation of an 800 V DC Power Supply for Next-Generation GPU Data Centers
Project Field: Electrical Engineering — Power Electronics, Digital Control, and Energy-Efficient AI Data-Center Infrastructure.
Project Description: This project investigates efficient and reliable power conversion for emerging 800 V DC distribution architectures in GPU data centers. Building on an existing converter design, students will contribute to the modelling, control, design verification, and experimental evaluation of an isolated laboratory prototype with a provisional power target of 1.5 kW and selectable 400 V or 800 V DC output.
The architecture combines a silicon carbide (SiC) totem-pole power-factor-correction (PFC) stage with an isolated dual-active-bridge (DAB) DC–DC converter, a high-frequency transformer, and Texas Instruments C2000 digital control. The input is single-phase 220–253 V AC at 60 Hz.
The central research question is how converter control can maintain stable output voltage under changing loads while improving efficiency and limiting component stress. Students will develop simulation models, investigate control strategies, review selected PCB and protection circuits, and compare simulated and experimental performance. Evaluation targets include efficiency of at least 95% and input power factor of at least 0.99 at nominal input and full load, output regulation within ±2%, and voltage ripple no greater than 1% peak-to-peak.
Expected outputs include documented simulation models, evaluated design and control improvements, and a final technical report and presentation. Experimental work will proceed through supervised testing after the documented hardware corrections and controller validation. The project focuses on the AC-to-DC supply stage feeding the 800 V bus.
Project Manager: Dr. Asif Islam asif.islam@kfupm.edu.sa
Project Title: Investigation into the Thermal and Dielectric Stress for Rated Load-current Switching in Medium Voltage Distribution Network
Project Field: High voltage testing, power system protection, electric arc modelling
Project Description: In recent year, there has been a considerable interest in SMART medium voltage (MV) electricity distribution networks, the practical implementation of which is limited by the availability of appropriate devices for MV switching and isolation. While SF6 interrupters are widely used in the distribution network for their exceptional switching capability, many countries are trying to reduce their usage considering the greenhouse properties of SF6 gas. Either developing a new type of environment-friendly circuit breaker or modifying an existing product to reduce SF6 usage, the next challenge is to type test the current-switching capacity against popular standards like IEEE, IEC, British Standard (BS) or Australian Standard (AS) etc. Switchgear manufacturers and type testing laboratories conduct the active load-current switching test of MV interrupter switches (circuit breaker, automatic circuit recloser, load break switch or disconnector) following the guidelines of either the IEEE or IEC standard. However, the guidelines defined by these two standards to calculate the test circuit parameters are different; therefore, the dielectric and thermal stress of IEEE and IEC test TRVs are different. TRV stands for transient recovery voltage. In this research, we will investigate the test TRV of load switching tests in IEEE and IEC standard to find out which standard apply higher electrical stress on the test object. Simulations will be done in PSCAD for 15, 27 and 38kV 100A and 800A load break switch. Two- and four-parameter TRV envelopes will be used to characterize the TRV. The impact of load power factor on the shape of the test TRV will also be examined for both standards. To establish the acceptance of the simulated results, the calculated rate-of-rise of TRV (RRTRV) from simulations will be compared with Jonsson’s experiments in early 2010s, and the critical RRTRV vs. rate-of-fall of current characteristics will be compared with Birtwhistle’s experiment of the early 1970s. The severity of the applied thermal stress by different standard tests will be examined comparing the initial time delay and RRTRV within the first hundred microseconds of the TRV. Finally, the dielectric stress will be investigated by comparing the highest peak of TRV and the corresponding time to reach the peak TRV. The outcome of this project will not only be useful to the local switchgear manufacturers like Alfanar, GEDAC Electric, Whatoop, MEMF Electric, Gulf Transpower etc. to get their products type tested against different standards but also will be informative for the Saudi type testing laboratories like GCC lab to optimize their scope of ISO 17025 testing accreditation. This project, in turn, will facilitate the objective of KFUPM to promote industry-oriented applied research.
Project Manager: Dr. Asif Islam asif.islam@kfupm.edu.sa
Project Title: Design and Development of an IoT-Cloud Integrated “Plug’N’Save” System for Energy Efficient Smart Buildings [INSE2637]
Project Field: Sustainable energy, IoT, Smart home systems, peer-to-peer communication
Project Description: The increasing electrification of residential environments and the rising demand for energy-efficient living underscore the need for advanced, appliance-level monitoring and intelligent control solutions. Traditional smart plugs provide basic on/off control and coarse energy readings but lack the accuracy, communication robustness, and intelligence required to support modern home energy management systems (HEMS) and national sustainability goals. To address these limitations, this project proposes Plug’n’Save, a next-generation IoT-integrated smart plug designed for real-time energy monitoring, resilient connectivity, distributed device-to-device coordination, and enhanced power quality (PQ) management at the plug level. Plug’n’Save combines precision sensing, embedded intelligence, and hybrid communication technologies to deliver a comprehensive, scalable solution for residential energy optimization. The system integrates an ESP32 microcontroller, ACS712 and ZMPT101B sensors, and an optocoupled relay within a custom-designed PCB, enabling accurate measurements of voltage, current, power, and energy consumption. A key innovation is the integration of dual Wi-Fi/Bluetooth Low Energy (BLE) connectivity, ensuring continuous operability even under weak or intermittent Wi-Fi conditions. Furthermore, Plug’n’Save supports smart plug–to–smart plug communication, enabling distributed decision-making and coordinated load control without relying solely on cloud infrastructure. The project will deliver a fully functional smart plug prototype, a mobile application, a cloud dashboard, a PQ management module, and technical documentation suitable for commercialization. Upon completion, Plug’n’Save will advance IoT-based residential energy management through improved accuracy, resilience, and intelligence. The system will provide measurable reductions in energy waste, support grid stability, enhance household electrical safety, and offer actionable insights into appliance-level energy consumption.
Project Manager: Dr. Asif Islam asif.islam@kfupm.edu.sa
Project Title: Solar–Battery Sizing and Grid Feasibility for Critical Facilities
Project Field: Solar photovoltaics (PV), battery energy storage systems (BESS)
Project Description: Hospitals, data centers, and cooling facilities require uninterrupted power during utility outages. While solar photovoltaics (PV) combined with battery energy storage systems (BESS) offer an alternative to diesel backup, project viability depends on two connected requirements. The first one is the Techno-Economic Sizing, where the minimum-cost PV and BESS capacity required to sustain critical loads over a target outage duration is identified. The other one is Network Feasibility, which deals with the local distribution network that can accommodate the interconnected capacity during grid-tied operation and support the facility when islanded. This project uses optimization tools (ReOpt/HOMER) to size a resilient PV-BESS installation using local economic parameters, then evaluates its impact on feeder voltage and line loading using power flow analysis (DIgSILENT PowerFactory or MATLAB).
Project Manager: Dr. Asif Islam asif.islam@kfupm.edu.sa
Project Title: Grid-Forming vs. Grid-Following Converter Stability in Low-Inertia Power Grids
Project Field: Power system analysis
Project Description: Traditional power systems rely on synchronous generators to establish system voltage and frequency while providing mechanical inertia. As inverter-based renewable resources displace synchronous generation—aligned with Saudi Arabia’s target of 130 GW of renewable capacity by 2030—system inertia drops significantly, creating weak-grid stability challenges. This transition highlights a critical operational difference between two primary converter control strategies. One is the Grid-Following (GFL) Control that tracks an established grid voltage and synchronizes using a Phase-Locked Loop (PLL). While effective in stiff networks, GFL converters experience instability under low short-circuit ratios (SCR). The other one is Grid-Forming (GFM) Control that regulates its own voltage magnitude and frequency, acting as a voltage source. While inherently suited for weak networks, its operational limits and comparative stability margins require precise quantification. This project uses dynamic RMS simulations (DIgSILENT PowerFactory or MATLAB) to systematically evaluate the penetration thresholds and stability boundaries of both control strategies under identical network disturbances.
Project Manager: Dr. Mohamed Ali Abido mabido@kfupm.edu.sa
Project Title: Improving Efficiency and Reliability for Next-Generation Propulsion Systems of EVs over Extended Speed Ranges with AI-Assisted Predictive Controls
Project Field: Electric Vehicles
Project Description: This project proposed a novel technique that relies on a model-free predictive control scheme that is assisted by artificial intelligence methods. The performance of the proposed technique is further improved by combining the features of sliding mode control (SMC), online parameter estimation schemes, and real-time fault detection methods. This combination can help develop an intelligent control system that can be adopted to control the permanent magnet synchronous machines (PMSMs) with excellent precision and robustness; it can also be extended to involve the control of induction machines (IMs) and switched reluctance machines (SRMs) as powerful candidates for future EVs.
Project Manager: Dr. Muhammad Khalid mkhalid@kfupm.edu.sa
Project Title: Development of a Cost-Efficient Wide-Voltage Bidirectional EV Charger with G2V and V2X Capability Using a Quasi-Single-Stage Resonant Topology
Project Field: EV Charging
Project Description: This project aims to develop a low-cost, efficient EV charger that can operate over a wide voltage range and support bidirectional power flow, allowing electricity to flow from the grid to the vehicle (G2V) and from the vehicle to external loads or the grid (V2X).The charger will use an advanced resonant converter design to reduce power losses, minimize circuit components, and improve overall efficiency.
Project Manager: Dr. Ahmar Shafi ahmar@kfupm.edu.sa
Project Title: Development of a VR-Based Interactive Tool for 4G/5G Cellular Network Planning and Coverage Analysis
Project Field: Wireless Communications, 4G/5G Cellular Networks, Simulations, Virtual Reality, and Engineering Education
Project Description: This project will develop a simple proof-of-concept Virtual Reality (VR) tool for visualizing cellular network planning concepts using VR Setup (headsets and controllers). The student team will use existing MATLAB-based 4G or 5G simulations and a 3D campus or urban map and implement the same using VR technology. Users of the system would be able to explore a small set of base-station configurations and adjust selected parameters, such as transceiver location, antenna height/angles, transmit power or etc. The tool will display the resulting predicted coverage and signal quality as visual overlays. Students will compare selected VR outputs with MATLAB results and prepare a short user guide. The core goal is to deliver one working VR scenario. AI-based comparison and recommendations to improve the design, would also be achieved if time permits but not part of original scope.
Project Manager: Dr. Brahim Brahmi brahim.brahmi@kfupm.edu.sa
Project Title: Development of Bio-Inspired Artificial Organs for Robotic Interaction and Surgical Training
Project Field: Robotics and Mechatronics, Biomedical Engineering, Artificial Intelligence and Control Systems, Sensors and Haptic Technologies
Project Description: This project aims to design and develop artificial organs that mimic the mechanical properties of human tissues, such as stiffness, elasticity, and deformation. The artificial organs will integrate sensors to measure forces, pressure, and deformation during physical interaction.
The developed platform will provide realistic force feedback for robotic systems and create practical training environments for surgeons. Potential applications include medical robotics, robotic-assisted procedures, and surgical simulation. The project will involve mechanical design, material selection, sensor integration, and experimental testing.
Project Manager: Dr. Marif Siddique marif.siddique@kfupm.edu.sa
Project Title: AI-Driven Design of High-Frequency Power Converters for Wireless Power Transfer Systems
Project Field: Power Electronics, Wireless Power Transfer (WPT), and Artificial Intelligence, Machine Learning Applications
Project Description: This project focuses on using computer simulations and artificial intelligence to study and improve the efficiency of wireless power transfer converters. Students will design and simulate a high-frequency resonant power converter using MATLAB/Simulink or PLECS. They will vary parameters such as input voltage, load resistance, switching frequency, duty ratio, and power level, and collect the simulation results. The collected data will be processed in Python and used to train a machine learning model that can predict converter efficiency, output voltage, current, and power losses under different operating conditions. The model will also help identify the best operating conditions and converter design parameters quickly without running a new simulation each time.
Project Manager: Dr. Marif Siddique marif.siddique@kfupm.edu.sa
Project Title: Design of a Power Converter for Wireless EV Charging Applications
Project Field: Power Electronics, Wireless Power Transfer
Project Description: This project focuses on designing and simulating a power converter for wireless electric vehicle (EV) charging using MATLAB/Simulink or PLECS. The system will include a high-frequency inverter, resonant circuit, wireless power transfer coils, rectifier, filter, and EV battery load. Students will select suitable circuit components and evaluate the converter under different operating parameters, including input voltage, charging power, battery voltage, switching frequency, coil distance, coupling coefficient, and coil alignment. The design will be assessed based on output voltage, charging current, efficiency, power losses, voltage and current stresses, and power quality.