Journal of Engineering and Sustainable Development (JEASD)

Generative artificial intelligence is advancing at a blistering pace. Large Language Models, in particular, have sped up the development of machine learning applications. This work presents a large language model-based technique to query data collected during MD2 pineapple crop production. Retrieval Augmented Generation was used to feed structured and unstructured data to two large language models (GPT-4 and LLAMA2) to train and fine-tune the models. The performance of the models was then measured using actual and predicted question-answer pairs. Results showed that the models had a 78% - 87% correct answer rate for structured and 75% - 79% correct answer rate for unstructured data. However, results showed that the models had a 61%-68 % correct answer rate when an answer to a question needed to refer to structured and unstructured data. These results showed that large language models can be further investigated to give farmers useful insights when making crop management decisions.

The risks of erosion to road embankments due to increased river water volume, particularly during flood events, often disrupt economic activities such as food and clothing distribution between locations. Therefore, this research investigated the critical impact of rapid drawdown on slope stability for road infrastructure adjacent to the Laeya River in Southeast Sulawesi, Indonesia. To achieve this aim, a comprehensive two-stage method was adopted, and the first included hydrological analysis to simulate flood water levels and predict rapid drawdown scenarios. In the second stage, a detailed geotechnical analysis through finite element method (FEM) was used to assess the stability of the slopes. Consequently, the research findings showed essential safety factor (SF) values under various conditions. The initial SF for the existing condition was 1.20, but after implementing treatment measures, such as slope geometry modification and soil compaction, the value improved to 1.54. However, during flood water level conditions, SF decreased to 1.50 due to the submergence of the slope base. Observation showed that rapid drawdown conditions led to a critical reduction in SF to 1.22, signifying the need for further research on the implications of drawdown. This research provided valuable perceptions and engineering solutions for improving slope stability in river-adjacent road infrastructure.

Photocatalytic reduction of CO2 to valuable chemicals and fuels requires highly efficient semiconductor materials, and most available photocatalysts are less efficient. Due to their unique electrical properties, the new family of two-dimensional materials known as MXenes has drawn much interest in photocatalytic applications. The vanadium carbide (V2C) is one of the significant MXenes being considered because of its many advantages over other materials. In this work, V2C-loaded TiO2 composites were synthesized and tested for photocatalytic reduction of CO2 with H2O to produce value-added CO and CH4 fuels in a continuous flow photoreactor system. The optimized 10 % V2C-TiO2 was responsible for CO and CH4 formation of 1233.8 and 85.2 µmol g-1 h-1, respectively, which were many-fold higher than using pure TiO2. This enhanced photoactivity of the composite was due to increased conductivity, many active sites, and higher light absorbance, which allowed for the efficient separation of charge carriers and light absorbance. Thus, MXenes, particularly 2D V2C MXene, would be a promising cocatalyst to combine with a semiconductor to maximize photocatalytic activity during CO2 reduction application.

This article proposes a linear Thevenin's equivalent circuit for a single-diode model. The Thevenin's equivalent circuit is derived based on the piecewise linearization for the nonlinear characteristic of the diode. To validate the accuracy and efficiency of the proposed Thevenin approximation, a photovoltaic module’s current-voltage and power-voltage characteristics are evaluated and compared with the characteristics of the original model. Meanwhile, an error of the proposed approximation is also calculated under different values of irradiance and temperature. The operating conditions are classified into three scenarios: standard technical condition (irradiance=1000 W/m2, temperature=25 oC), (irradiance=400 W/m2, temperature=25 oC), and (irradiance=1000 W/m2, temperature=50 oC). Primary results show encouraging outputs and reliability of the proposed model. It is obvious from the simulation results that the linearization technique can successfully emulate the characteristics of the original nonlinear single-diode model. The peak values of the error are 7.37%, 9.51%, and 6.79% at the different suggested scenarios, respectively. Subsequently, the proposed Thevenin equivalent circuit can be successfully used to study the performance of a photovoltaic system under different operating conditions, avoiding the complicated numerical solutions for nonlinear equations of the original photovoltaic model.

This paper uses a finite control set (FCS) model predictive control (MPC) technique to control the output voltage of a 3-phase 4-leg multilevel inverter with a minimum number of switching states to supply a balanced, unbalanced, or nonlinear load. Good results are obtained when using this technique for a flying-capacitor (FC) multilevel inverter with an output that includes a filter made of capacitors and inductors. In this method, 19 switching states are used instead of 81 switching states in classical inverters, reducing the time required to reach the optimal solution in the cost function (g). The cost function calculates the minimum error between the output and reference voltage by selecting the optimal switching state for driving all inverter transistors. This technique can be applied to other types of multilevel inverters without considering the type of load used or needing to adjust the controller parameters. The inverter is designed for powering the 3-phase loads by separately generating three output voltages from the fourth leg using MPC

This paper proposes the optimization of planar antenna arrays for DOA estimation enhancement in radar systems, where accuracy and computational efficiency need to be well balanced. This work describes a new methodology using GAs to obtain an optimal configuration of the antenna elements, allowing a considerable reduction in computational complexity. Using the ESPRIT algorithm for azimuth and elevation angle estimation, the proposed approach achieves an RMSE as low as 0.2 degrees under an SNR of 20 dB. For practical setups, an antenna size configuration of M = 8, N = 3 with 85 snapshots can achieve an optimal tradeoff, reducing the computational cost by 35% compared to conventional methods while maintaining high accuracy. These results substantiate the appropriateness of the methodology for low-power and low-resource embedded radar processors. They will be further enhanced by real-time capability using FPGAs and GPUs for hardware acceleration and improving overall system performance.

Many industries use robotic arms to perform tasks like picking and placing. The main goal of this paper is to derive and implement the forward and inverse kinematics of a 6-degree-of-freedom articulated robotic arm. In addition, a closed-form solution is required to design the inverse kinematics, unlike the forward kinematics. Then, the inverse kinematics and forward kinematics modeling were successfully performed on the WidowX 250 6-degree-of-freedom robotic arm. The MATLAB modeling results are compared to the ROS results, which showed a matching percentage of 99%. In addition, the inverse kinematics results revealed a range of solutions that are approximately identified to the desired results, while others were not, such as the value of the third joint angle is obtained at ±1.57 radians. Finally, due to the accuracy value of this comparative work between real ROS and MATLAB derived equations of the WidowX 250 6DOF robot arm, it is concluded that the joints and coordinates of this robot are identified.

The idea is to use alternative materials related to materials with functional classification and porosity in the human body. For example, the need for a replacement for human bones should be lightweight, rigid, non-toxic, and have an internal porosity that differs from the external surface. People now live longer and more comfortably when they experience fractures, amputations, and vertebral replacements, bone implants, which were previously unavailable or were limited and very expensive. These advancements highlight the ongoing efforts to enhance materials used in medical contexts, emphasizing functionality, biocompatibility, and the ability to interact harmoniously with the human body. The idea is to use alternative materials related to materials with functional classification and porosity in the human body. For example, the need for a replacement for human bones should be lightweight, rigid, non-toxic, and have an internal porosity that differs from the external surface. People now live longer and more comfortably when they experience fractures, amputations, and vertebral replacements, bone implants, which were previously unavailable or were limited and very expensive. These advancements highlight the ongoing efforts to enhance materials used in medical contexts, emphasizing functionality, biocompatibility, and the ability to interact harmoniously with the human body. The present study provides a review of materials with functional classification and porosity in medical applications and methods of their manufacture, which can be successful substitutes in the human skeletal system. Recently, regarding biocompatibility, PLA is considered excellent, making it suitable for medical and biomedical applications. The use of 3D printing technologies in manufacturing porous structures from various materials such as PLA, ABS, and Titanium shows that these manufactured structures have excellent mechanical properties and good biocompatibility. Based on the results, it is possible to design customized structures that meet specific needs by changing the size, shape, and porous structure ratio.

In prosthetics, 3D scanning technology plays a crucial role in gathering data for designing prosthetic devices. New tools for capturing the external shape of human body parts are rapidly emerging. This study conducted a comparative analysis to assess the accuracy and precision of two handheld 3D scanners, including the Peel 3 3D scanner and a smartphone application, in scanning the residual limb of a transtibial participant. The residual limb was manually measured, and six scans were taken with each scanning system, compared to the reference computed tomography (CT) model. A repetition test was performed to determine the mean value and standard deviation of the scanned models in VXelements software for comparison purposes. The results showed a standard deviation difference between the Peel 3 and CT scans ranging from ± 0.828 to 0.907mm, and from ± 0.823 to ± 0.877mm for the smartphone scans. The repetition test indicated standard deviations ranging from ± 0.531 to 0.599 mm for the Peel 3 and ± 0.458 to 0.690 mm for the smartphone. This analysis evaluated the accuracy of the 3D scanners and identified the essential specifications for their effective use in prosthetics.

Low-emissivity (low-e) glass is a highly efficient material capable of adapting its thermal and optical properties in response to temperature fluctuations without consuming additional Energy. This makes it an ideal choice for passive building envelopes, particularly in climates with extreme temperatures. In this study, the thermal efficiency of triple-glazed windows incorporating low-e glass was experimentally evaluated under the harsh summer conditions of Iraq. Specifically, a conventional triple-glazed window (TG1) with air-filled gaps was compared to a modified version (TG2) featuring low-e glass on the outer pane, also with air in both gaps. The research findings revealed that when solar radiation peaked at 650 W/m² around midday in July, the internal surface temperatures of TG1 and TG2 were 36.4°C and 32.4°C, respectively. The temperature reduction observed in TG2, amounting to 4°C (10.9%), highlights the significant thermal advantage the low-e glass provides, making it a superior option for enhancing energy efficiency in buildings.

In Iraq, where electric generation is limited, a pilot work was conducted at East Karbala Diesel Power Station, in Karbala. This aimed to assess the implementation of PV solar systems on stations under both grid-connected and off-grid conditions. We will evaluate various factors, including net present cost, energy cost, operating and initial costs, and annual net energy consumption from the grid. The work analyzed various situations and considered the presence of backup generators (G1-250kVA) at the station. The first step is to conduct a walkthrough energy audit at EKDPS to determine the plant's load profile. By using HOMER Pro, further analysis is illustrated. Results indicated that integrating an on-grid PV system could reduce the cost of energy by 58%, making it the optimal situation when grid power is available. In this case, the Net Present Cost, operating cost per year, was $51,991. Conversely, in the absence of grid power, the best scenario involved a PV system with a 100-kWh lithium battery loading. Despite being the most effective resolution in this situation, the energy cost was 310% greater than the one-case energy rate in Iraq (0.1 $/kWh). Notably, both optimal scenarios did not require generators, highlighting a significant potential benefit for the Nero Zero system.

An emphasis is placed on the optimal location and sizing of distributed generation and reactive power compensation to minimize real power loss in the Radial Distribution Network. This paper uses the Reconfiguration Method to determine the optimal locations of reactive power compensators and distributed generation units. Grey Wolf Optimizer (GWO) is the technique that is used to select the optimal sizing of reactive power compensators and distributed generation units. Two approaches are used: the meta-heuristics optimization technique GWO, and the Reconfiguration Method to determine the optimal location and size of shunt capacitors and distribution generators. This paper aims to minimize the real power losses, improve the voltage profiles, and enhance the voltage stability index on distribution networks by adding distributed generators and shunt capacitors with the optimal size. The IEEE 33-bus system, as a standard power system, is utilized to validate the suggested technique. Simulation results with various case studies show that the placement of distributed generating units and reactive power compensators in power distribution networks reduces the stress on the line loads and produces a sizable loss reduction with a favorable voltage profile.

The three ultrafiltration membranes prepared from different concentrations (13, 15, and 17 wt.%) of Polyvinylidene fluoride polymer are used to remove Congo Red Dye. The phase inversion method is utilized to fabricate the membranes. Membrane characterizations were studied using atomic force microscopy (AFM) to examine roughness. Scanning electron microscopy (SEM) was used to compare morphology and cross-sectional structure. Energy-dispersive spectrometry (EDS) determines the number, type, and distribution of atoms. Results showed that increasing the polymer concentration to 17 wt.% enhanced the membrane's hydrophobicity and improved dye rejection, attributed to higher surface roughness, a uniform and dense distribution of chemical elements (higher F and C peaks), and reduced porosity. Membrane performance was studied under operating pressures (2-7 bar) using a 100 ppm Congo Red dye solution in a cross-flow filtration system. The PVDF 17 wt.% membrane showed a flux rate of 4.46–41.89 L/m²·h at 2 and 7 bar, respectively, with 99% dye removal at 2 bar, decreasing to 94% at 7 bar.

Zinc, an essential element, offers numerous benefits; however, an excessive accumulation of zinc within the human body can pose detrimental effects. This study endeavors to mitigate zinc contamination in water sources through an adsorption process utilizing buckthorn leaves as a cost-effective, natural adsorbent. The investigation examined various operational parameters that affect the adsorption process, including pH levels, initial zinc ion concentrations, buckthorn leaf dosage, contact duration, temperature, and agitation speed. The results showed that 78% was the maximum removal of zinc ions, achieved at the optimum conditions of 6, 25 ppm, 4.5 g, 120 min, 45 °C, and 350 rpm, respectively. Furthermore, the residual buckthorn leaves, laden with zinc ions, were evaluated for potential utilization as an organic fertilizer for white beans. The investigation entailed determining the zinc ion content adsorbed by the leaves. Subsequent cultivation trials demonstrated a noteworthy enhancement in bean pod yield from soil treated with zinc-loaded leaves, exhibiting a 9% increase compared to soil treated solely with raw leaves or left untreated. Notably, the resulting crop mass was equivalent across all treatment groups. Thus, this study presents an innovative approach to simultaneously manage solid waste and purify water, offering an efficient, cost-effective, and environmentally sustainable solution

In this paper, an ultra-wideband terahertz monopole antenna is proposed to address the limitations of lower bandwidth and low gain in previously designed antennas at THz frequencies. The antenna is composed of three parts: a pentagonal patch, a modified ground plane that provides impedance matching (2.5-12) THz, and an FR-4 material substrate. The complete proposed antenna has been simulated using electromagnetic software and a computer simulation tool. Based on the simulation findings, the proposed antenna is the best option for Tera-Hertz applications. The antenna has an acceptable gain in the intended frequency spectrum band and operates between 2.4 and 12 THz. Computer simulation technology (CST 2020) software has been used to Analyze and simulate the proposed antenna. The simulation results show that the return losses of the frequency band (2.4-12) THz are less than -10 dB, and the gain has an appropriate value over the frequency band 2.4-12) THz. The radiation pattern of the antenna is omni-direction at the desired frequency band.