Al-Khwarizmi Engineering Journal

Biodiesel is becoming one of the most attractive alternative biofuels for petroleum-based diesel fuels. The castor plant is one of the abundant non-edible oils found in many countries. This paper aims to study Libyan castor oil and its potential for diesel conversion. Experiments were carried out in the laboratories of the Specific Center for Training in the Oil Industries in Al-Zawiya. The oil was extracted using a Soxhlet extractor and n-hexane solvent at 60 °C. Transesterification reactions were conducted in a batch reactor (a three-neck flask was used, where the middle opening carries a reflux condensation unit) at 65 °C. The methanol-to-castor oil molar ratio was 6:1, with a catalyst concentration of 1 wt.% relative to the castor oil, and the reaction time was 30 min. Castor oil was analysed and found to have a fatty acid content of 0.7%. The productivity of biodiesel exhibited 80% yield. Standard test methods of analysis were conducted to determine the biodiesel’s properties. Results indicated that the kinematic viscosity at 40 °C was 14.24 cSt, the density was 0.924 g/cm3, the cetane number (CN) was 54 and the pour point was −15 °C, indicating that castor oil biodiesel is of high quality, with a high CN and a low pour point.

Image segmentation has a significant role in the virtual planning, execution, and evaluation of craniomaxillofacial (CMF) surgical procedures. This systematic review aims to evaluate and compare the image segmentation programs frequently used in the field of CMF surgery. A precise search strategy was employed to recognise suitable studies across several databases, using specific inclusion criteria and keywords. Various image segmentation programs that use different techniques, including thresholding, edge-based methods, region-based methods and machine learning-based methods, were investigated. Results were screened through Preferred Reporting Items for Systematic Reviews and Meta-Analyses. A total of 94 reports on the use of virtual surgical planning from January 1, 2014, to June 1, 2023, were obtained. The identified image segmentation programs were analysed, including factors such as program features, strengths, limitations, supported image modalities, and clinical applications. A qualified assessment of these programs was conducted on the basis of parameters such as segmentation accuracy, processing speed, robustness, user-friendliness and integration capabilities. The review also addresses challenges faced by current segmentation programs and outlines future directions for advancement, including the standardised validation metrics and the integration of artificial intelligence. Surgical procedures were assigned into seven categories for analysis: cranial reconstructions, facial rejuvenation, orthognathic surgery, trauma repair, tumour resection, cleft lip and palate and patient specific implant. Amongst the software that could be used for bone segmentation in CMF region, eight software programs are most frequently used. Results showed that the Materialise suite was the most widespread tool for bone segmentation programs, with a prevalence of 50%, followed by the 3D slicer. This review underlines the principal significance of image segmentation in CMF surgery and offers valuable insights for clinicians and researchers to make informed decisions regarding the selection and utilisation of image segmentation programs.

This study investigates the use of agricultural waste, specifically the Al Haji plant, for wastewater treatment. The term \\'Al Haji plant\\' refers to species from the Al Haji genus, particularly \\"Al Haji maurorum \\" and \\"Al Haji canescens \\", which are drought-resistant plants of the Fabaceae family. Traditionally, these plants are known for their medicinal and antimicrobial properties. In this research, waste from the Al Haji plant was converted into activated carbon (AC) as a low-cost and techno-economically viable biosorbent for the removal of methylene blue (MB) dye from wastewater. The production of AC involved two key steps: chemical activation using sodium hydroxide (NaOH) and carbonisation at 300 °C. The properties and adsorption potential of the resulting AC were investigated. Various parameters were studied, including pH, contact time, stirring rate, initial dye concentration, adsorbent dosage, and temperature. Characterisation of the biosorbents was performed using Scanning Electron Microscopy (SEM) and Brunauer–Emmett–Teller (BET) surface area analysis. The SEM investigation showed notable pores in the Al Haji plant sample that could enhance the absorption of the MB dye. Results indicated that the BET surface area for physical AC is 25.5796 m²/g. The surface area of chemical AC was reported as 16.2767 m²/g. The dye exhibited pseudofirst-order adsorption kinetics, with average rate constants of 0.020082 and 0.001589 for physical and chemical activation approaches, respectively. An adsorption isotherm study was performed using the Langmuir model, which showed an R² value of 0.827 and a maximum adsorption capacity of 5.24 mg/g.\\" The study of adsorption thermodynamics revealed that MB adsorption on Al Haji activated carbon is spontaneous and endothermic.

The aim of this study was to implement and compare obstacle avoidance for an autonomous wheeled mobile robot (WMR) via the grey wolf optimisation (GWO) algorithm and the artificial bee colony (ABC) algorithm. The study was conducted via three scenarios, each designed to test the performance of the algorithm under different conditions, considering fixed and moving circular obstacles in the surrounding environment. GWO was used to determine the most efficient, shortest and safest path for the WMR from the starting point to the target point. The results showed that the GWO outperformed the ABC. The GWO also enabled the WMR to avoid obstacles faster by 11.8%, 2.8% and 4.6% and with distances shorter by 1.42%, 2.2% and 1.97% for the three scenarios, respectively.

Industrial automation has significantly enhanced productivity, efficiency, and product quality by integrating advanced technologies such as robotics, artificial intelligence, and IoT. This study focuses on rehabilitating and developing the DL-MS6000 automatic industrial production line training system at the Ministry of Higher Education and Scientific Research in Baghdad, Iraq. The system, comprising six stations controlled by six PLCs, had issues, including hardware failures, the need for periodic updates, and a lack of SCADA control. The development process replaced the six PLCs with a single CYMON PLC and eliminated the need for CC-Link network devices, reducing system complexity and cost. The system was optimized using a new control algorithm. The system achieved high precision, zero steady-state error, and a settling time of 40 ms by using the Root Locus method in the MATLAB program. Additionally, a SCADA/HMI system was implemented to enhance control and monitoring. The newly developed system offers increased reliability, cost-efficiency, and ease of use, making it a suitable tool for educational purposes in university training rooms.

Drones play a vital role in the fundamental aspects of Industry 4.0 by converting conventional warehouses into intelligent ones, particularly in the realm of barcode scanning. Various potential issues frequently arise during barcode scanning by drones, specifically when the drone camera has difficulty obtaining distinct images due to certain factors, such as distance, capturing the image whilst flying, noise in the environment and different barcode dimensions. In addressing these challenges, this study proposes an approach that combines a proportional–integral–derivative (PID) controller with image processing techniques. The PID controller is responsible for continuously monitoring the camera’s input, detecting the difference between the planned and the real barcode image dimensions, and making immediate changes to the drone position to improve the process of detecting the potential barcode. The aforementioned procedure is implemented on the DJI Tello drone to verify the practical performance of the methodology introduced in this study. Results showed that drones can achieve remarkable barcode scanning performance by incorporating sophisticated computer vision technologies into PID controllers. PID computer vision algorithms are capable of analysing visual data acquired from the drone’s cameras and retrieving barcode information under a variety of situations, such as the size of the barcode, location of the barcode and noise of the warehouse environment.

Software-defined networking (SDN) has proven its superiority in addressing ordinary network problems, such as scalability, agility and security. This advantage of SDN comes because of its separation of the control plane from the data plane. Although many studies have focused on SDN management, monitoring, control and improving quality of service, only a few are focused on presenting what is used to generate traffic and collect data. The literature also lacks comparisons amongst the tools and methods used in this context. Therefore, this study introduces the recent tools used to simulate, generate and obtain traffic statistics from an SDN environment. In addition, the methods used in SDN data gathering are compared to explore the capability of each one and hence, determine the suitable environment for each method. The SDN testbed is simulated using Mininet software with tree topology and OpenFlow switches. An RYU controller was connected to control forwarding. The famous tools iperf3, Ping and python scripts are used to generate network datasets from selected devices in the network. Wireshark, the RYU application and the ovs-ofctl command are used to monitor and gather the dataset. Results show success in generating several types of network metrics to be used in the future for training machine or deep learning algorithms. Therefore, when generating data for the purpose of congestion control, iperf3 is the best tool, whilst Ping is useful when generating data for the purpose of detecting distributed denial-of-service attacks. RYU applications are the most suitable monitoring tool for obtaining all network details, such as the topology, characteristics and statistics of the components. Many obstacles and mistakes are also explored and listed to be prevented when researchers try to generate such datasets in their next scientific efforts.

This study aims to improve the performance of reconfigurable STAR-passive-OIRS in multiple input multiple output-based free-space optical (FSO) communication systems. The study focuses on analysing the influence of two main factors: the number of reflective elements and the location of the smart surface. STAR-RIS technology has shown great potential in improving spectrum efficiency and signal-to-noise and interference ratio (SINR) and in increasing channel gain by providing 360° coverage and reducing interference in environments with line-of-sight constraints. Results reveal that increasing the number of reflective elements enhances signal strength and improves SINR, whilst placing the smart surface closer to the source or receiver achieves optimal performance. The study highlights STAR-RIS technology’s ability to overcome traditional FSO systems’ limitations and improve network performance in modern wireless communication environments.

Phishing detection is considered a critical problem in cybersecurity, and utilising machine learning with an efficient feature selection method for precisely identifying malicious websites is deemed the most critical challenge. This research presents a two-phase phishing detection system by employing unsupervised feature selection and supervised classification. In the first phase, the best set of features is identified by the Genetic algorithm and is utilised by the K-means clustering algorithm to divide the dataset into groups with similar traits. In the second phase, the best set of features in each group is identified through the Genetic algorithm to enhance the classification process. Finally, a voting ensemble technique is applied, in which the Support Vector Machine (SVM), Random Forest (RF), Extreme Gradient Boosting (XGBoost) and Adaptive boosting (AdaBoost) models are combined. Predictions are aggregated using a soft voting mechanism. This research utilises the web page phishing detection dataset, which consists of 11,430 URLs with 87 features. From the results, an accuracy of 99% is achieved using the voting ensemble technique with feature selection compared with 77.3% without feature selection. The model performance experiences a significant boost through the GA-optimised feature selection by reducing computational complexity and improving key metrics such as accuracy, precision and F1-score. Additionally, the performance across four clusters demonstrates the positive impact of K-Means clustering in improving classification accuracy for specific data groups. As proven by the obtained results, integrating feature selection with ensemble learning is effective for phishing detection; moreover, the scalability and efficiency of such a solution in real-world applications are demonstrated.