Al-Qadisiyah Journal for Engineering Sciences

Abstract This study evaluates changes in hardness and estimates the reduction in tensile strength of BJ37 structural steel after exposure to elevated temperatures of 250 ^°C and 500 ^°C using the Leeb hardness non-destructive testing (NDT) method. Since hardness is closely related to mechanical properties, the Leeb test can be used as an indirect approach to estimate tensile strength without damaging the material. A total of 36 specimens were prepared from the flange and web sections of a WF 300 X 150 X 6 X9 steel profile. The specimens were heated to 250 ^°C and 500 ^°C for 15 minutes and then rapidly cooled by water immersion before testing. The initial hardness was measured at approximately 366.19 HL, corresponding to an estimated tensile strength of about 372 MPa. After heating to 250 ^°C, the hardness decreased to 352.5 HL, with an estimated tensile strength of 342.5 MPa; at 500 ^°C, the hardness was 354.7 HL, corresponding to approximately 346 MPa. The reduction in tensile strength ranged from 7.0% to 7.9% relative to the initial condition, with only a small difference between the two temperature levels. These results indicate that the Leeb hardness method can serve as a rapid, non-destructive tool for the preliminary assessment of steel mechanical properties after fire exposure. Further research is recommended to validate these findings through destructive tensile testing, considering a wider range of temperatures, longer heating durations, and a larger number of specimens. The findings provide a practical basis for rapid post-fire evaluation of steel structural integrity.

Abstract Advances in smart materials have made it possible to develop shape-changing wing concepts as alternatives to traditional high-lift systems. In this study, the aerodynamic effects of linear beam extension on a NACA 0015 wing profile with an initial beam length of 150 mm were numerically investigated. The linear actuation mechanism applied along the chord direction resulted in a reduction in the camber ratio. The aerodynamic performance of the conformable configurations was evaluated using CFD analyses conducted under low Reynolds number conditions and employing the k- SST turbulence model. The results indicate that chordwise deformation can improve aerodynamic performance under various operating conditions. This approach offers significant potential for enhancing maneuverability and reducing energy consumption in low-Reynolds-number aircraft, presenting a promising solution for future adaptive wing designs.

Abstract Cancer is still a major health concern, particularly in areas like Iraq with inadequate healthcare systems, where survival rates depend on early and precise diagnosis. Using clinical text data from radiology reports in Mosul, Iraq, this study examines the use of Natural Language Processing (NLP) and Machine Learning (ML) models for cancer diagnosis and classification. In order to categories cancer cases into benign, malignant, stable, progress, and improvement groups, three machine learning classifiers—Support Vector Machine (SVM), XGBoost, and LightGBM—were trained using TF-IDF features on a balanced dataset of 12,923 labelled radiological reports. XGBoost outperformed the other models and showed the highest accuracy (97.25%). This study examines the useful implications for improving diagnostic efficiency and demonstrates the efficacy of NLP-driven machine learning models in healthcare settings with limited resources. The results imply that these ML-NLP models can increase accuracy, decrease the need for manual diagnostic procedures, and possibly offer a scalable solution for healthcare systems with limited funding.

Abstract Natural convection heat transfer in corrugated enclosures has attracted increasing attention due to its importance in thermal management applications. However, the combined effects of wall corrugation, porous media, ethylene glycol–based nanofluids, and magnetic field control have not been sufficiently addressed. In the present study, a numerical investigation is carried out to analyze thermo-magnetohydrodynamic natural convection inside a corrugated enclosure subjected to different heating configurations. The enclosure is divided into two regions: an upper region filled with an ethylene glycol–based nanofluid and a lower porous layer saturated with the same nanofluid. Three heating cases are considered, namely an externally heated corrugated cylinder, an upper heated wall, and an externally heated circular boundary. The governing equations are solved using the finite element method under steady-state conditions. The effects of Rayleigh number (10⁴  up to 10⁶), Hartmann number (0 up to 60), Darcy number (10^-5  up to  10^-3), and magnetic field inclination angle (0° up to 35^o ) are examined at a fixed nanoparticle volume fraction of 0.02. The results show that increasing the Rayleigh number enhances heat transfer, while the magnetic field intensity suppresses convection. Among all cases, the externally heated corrugated cylinder provides the highest thermal performance, and the ethylene glycol–based nanofluid enhances the average Nusselt number by up to 48% at Ra = 10 ⁶.

Abstract This work examines notch sensitivity in fiber-reinforced polymer composites, with attention to the role of fiber orientation and notch width. CFRP and GFRP specimens were fabricated by vacuum-assisted resin transfer molding (VARTM) using thermoset resin systems. The tested variables were fiber orientation (0^o and 90^o), notch width (0, 2, 4, and 6 mm), and tensile tests were carried out according to ASTM D638. The results showed that the 0° specimens had noticeably higher strength than the 90^o specimens because the fibers were aligned with the loading direction. The 90^o specimens showed lower strength and a more gradual failure response, mainly due to matrix-dominated behavior. Increasing the notch width reduced the ultimate tensile strength, especially for 4 and 6 mm notches, where fracture started near the notch tip because of stress concentration. Unnotched and 2 mm notched specimens provided the best balance between strength retention and practical manufacturability. These results confirm that careful fiber alignment and notch control are important for composite parts used in aerospace, automotive, and civil engineering structures

Abstract This paper presents a technical and environmental assessment of a 3 kWp grid-connected residential rooftop PV system in Al-Diwaniyah, Iraq, using PVsyst 8.0.19. Four scenarios were analysed: monofacial and bifacial PV panels with and without a 5 kWh LiFePO4 battery, plus an expansion sensitivity case using a 10 kWh battery. The battery-free system generated 6,444.8 kWh/year, with a performance ratio (PR) of 80.72%, specific yield of 1,790 kWh/kWp/year, solar fraction of 64.37%, and grid imports of 1,262.2 kWh/year. Adding a 5 kWh battery slightly reduced PR to 77.86% due to charge/discharge losses but increased the solar fraction to 93.00% and reduced grid imports by 80.4% to 247.98 kWh/year. Bifacial panels improved annual production by about 90 kWh/year (+1.4%) and increased PR by 1.13 percentage points in both storage and non-storage cases. Expanding storage to 10 kWh raised the solar fraction only from 93.0% to 96.7%, showing diminishing returns and doubling the cost. Environmentally, the system avoids about 118.4 tCO₂ over 25 years, equivalent to 5,381 trees or removing 25.7 cars annually, with an avoided carbon cost of USD 5,920. The 5 kWh monofacial battery system is the optimal configuration for residential energy independence in the Iraqi context.

Abstract Lake Maninjau, West Sumatra, Indonesia, is under pressure from floating net cage aquaculture, shoreline settlements, agricultural runoff, and domestic activities. This study mapped the spatial distribution of pH and Total Suspended Solids (TSS) using an integrated Geographic Information System (GIS), Ordinary Kriging, and Sentinel-2 MultiSpectral Instrument (MSI) Level-2A imagery. Field sampling was conducted at 30 sites on 18 October 2025, and a cloud-free Sentinel-2 image acquired on 20 October 2025 was processed through Google Earth Engine. The Normalized Difference Turbidity Index (NDTI) and Normalized Difference Water Index (NDWI) were extracted to support spatial interpretation. Measured pH ranged from 5.8 to 8.2, while TSS ranged from 10 to 75 mg/L. Low pH values were concentrated near shoreline settlements and aquaculture areas, whereas high TSS values occurred near agricultural inflows and floating net cage zones. Cross-validation showed acceptable pH kriging performance, with RMSE = 0.896, MAE = 0.628, and R² = 0.755. TSS kriging showed weaker performance, with RMSE = 57.17 mg/L, MAE = 38.94 mg/L, and R² = -0.018. Integration with Sentinel-2 indices showed a strong relationship between TSS and NDTI (R² = 0.84, RMSE = 6.2 mg/L, r = 0.92, p < 0.01), while pH and NDWI showed a weaker relationship (R² = 0.69, r = 0.74, p < 0.05). The results indicate that Sentinel-2 imagery is more effective for optically active parameters such as TSS than for chemical parameters such as pH

Abstract In recent years, drone technology has become very significant in many industrial applications especially for civil infrastructure inspection. In this paper, we present a cascade Proportional Integral Derivative (PID) controller supported by an Artificial Intelligence (AI)-based defect detection system by using You Only Look Once (YOLOv8) algorithm using a drone for civil infrastructure. In conventional inspections, the inspectors are facing difficulties in identifying defects because of old and hazardous buildings, unseen small defects and having substantial safety risks while inspecting high-risk building or bridge structures. This study developed a low-cost Do It Yourself (DIY) drone which was outfitted with a flight controller and First Person View (FPV) camera to capture image in real-world scenarios. The developed AI-based DIY drone system utilized a high-resolution imaging camera which enables rapid and remote data acquisition. During flight testing, the camera captured 60 frames per second of real-time footage while maintaining stable navigation and carrying a payload of up to 800 grammes. The YOLOv8 model trained, validated and tested 6,998 annotated photos, which showed a high mean Average Precision (mAP@0.5) of 99.4%, precision of 96.9%, and recall of 98.4%. The results ascertained that the system can reliably identify structural flaws like cracks, corrosion, peeling paint, and water seepage in difficult environmental conditions.

Abstract The most lethal form of skin cancer is malignant melanoma; however, it is highly curable when detected at an early stage. Despite the exceptional diagnostic capabilities of Deep Convolutional Neural Networks (CNNs), it is essential to select the appropriate architectural paradigm in order to achieve the highest clinical safety, the lowest computation cost, and the best predictive performance. This investigation comprehensively evaluates the binary classification of dermoscopic images using five distinct pre-trained CNNs (SqueezeNet, GoogleNet, ResNet-50, MobileNetV2, and EfficientNetB0) and a variety of structural paradigms. A balanced dataset of 9,605 images was utilized alongside transfer learning, stochastic spatial augmentations, and an optimized Adam-based training protocol. To achieve consistent predictive stability (Accuracy = 90.94% ± 0.13%), EfficientNetB0 was tested using a 5-Fold Cross-Validation protocol, and the rest of the networks were tested using hold-out method with minimal computational burden. These findings of the quantitative analysis have illustrated that the clinical compromises of the architectures were diverse. The Area Under the Curve (AUC) and Specificity (98.40%) values of ResNet-50 were the highest, suggesting that it effectively reduced false alarms. On the other hand, GoogleNet, by emphasizing patient safety, obtained the highest Sensitivity (90.80%). Additionally, the SqueezeNet micro-architecture was capable of preventing class-collapse as a result of the optimized training process, which resulted in an increase in accuracy to 91.40%. This was critically important. The findings of this study indicate that deep residual networks are the most appropriate for high-precision screening, while the deployment of lightweight networks of a suitable scale is highly viable and reliable in computational environments with limited resources.

Abstract This study experimentally investigates the effect of hydrogen enrichment on combustion characteristics, engine performance, and exhaust emissions of a four-stroke spark-ignition (SI) gasoline engine. A single-cylinder, 0.61-L, air-cooled engine was operated at 2300 rpm under full load with hydrogen blending ratios of 5%, 10%, 15%, and 20% by fuel energy, using unleaded gasoline (RON 95) as the base fuel. Hydrogen was delivered through a dedicated port injector upstream of the intake valve. Peak cylinder pressure increased by 17.6% (4.82 to 5.67 MPa at H20%), and the 10–90% mass-fraction-burn duration decreased by 31.6% (38 to 26 °CA), driven by hydrogen’s high laminar flame speed (237 cm/s) and wide flammability range (4–75 vol%). Brake thermal efficiency peaked at H15% (32.1%, +12.6%), while brake-specific fuel consumption fell by 17.35% at H20%. Combustion stability improved by 60.7%, with COV of IMEP reduced from 2.8% to 1.1%. CO, HC, and CO₂ emissions decreased by 81%, 66.5%, and 25%, respectively, with the average unburned-HC carbon chain length reduced from C₆.₂ to C₃.₁. However, NOx emissions rose by 338.7% due to higher peak combustion temperatures activating the thermal (Zeldovich) mechanism. The optimal blending ratio is 10–15%, balancing efficiency gains against NOx penalties addressable through established aftertreatment strategies.