Journal of Engineering and Technological Sciences

A Comparative Study of Advanced Machine Learning and Deep Learning Models for Municipal Solid Waste Forecasting: A Case Study of Surat, India

2026-03-30T16:35:51+07:00

To make the city planning sustainable, especially in rapidly growing cities of the world like Surat in India, the implementation of effective waste management is crucial. The primary factor governing this is the ability to accurately predict the quantity of municipal solid waste (MSW) likely to be generated. This study presents a comprehensive comparative analysis of various predictive models including linear regressions, kernel approaches, gradient boosting as well as the deep learning architectures. Using historical data from Surat, systematic preprocessing and feature engineering generated 419 features representing temporal, socio-economic, climatic, COVID-19, and mobility factors. The novel contribution of this study is the systematic feature engineering framework that explicitly encodes temporal structure (419 engineered features including lagged values, rolling statistics, and seasonal decomposition), enabling simple linear models to capture complex waste generation patterns. Ten distinct models, ranging from statistical approaches to machine learning and deep learning were evaluated and compared. Advanced ensemble models, including LightGBM (R² = 0.983), CatBoost (R² = 0.977), and XGBoost (R² = 0.970) demonstrated strong performance. The best-performing models (OLS and Gaussian Process Regression) achieved R² = 0.997 with Mean Absolute Percentage Error (MAPE) = 1.43%. In this study, linear models trained within a few milliseconds and achieved per-sample inference times on the order of 0.004-0.008 ms, whereas the tuned MLP and tree-based ensembles required seconds of training and millisecond-level inference, corresponding to differences of roughly two to three orders of magnitude in computational cost. Other notable performers include Lasso regression (R² = 0.979), tuned MLP (R² = 0.968), and Random Forest (R² = 0.960). The results demonstrate that feature engineering has greater influence on forecasting accuracy than model complexity, with OLS using engineered features (R² = 0.997) outperforming the MLP model (R² = 0.966) by approximately 3.1% while providing substantially faster predictions. Feature importance analysis identified lagged MSW values, rolling statistics, demographic indicators, festival effects, and COVID-19 lockdown impact as key predictors. The research finds that adoption of systematically designed feature engineering framework is a valuable tool for MSW management. The study provides comprehensive model benchmarking and practical recommendations for policymakers pursuing sustainable urban development.

A Comparative Study of Snyder and Clark Hydrological Procedures for Flood Peaks Estimation

2026-03-05T08:51:05+07:00

The Snyder (HP11) and Clark (HP27) hydrological procedures are valuable tools in managing water resources in urban and rural areas across Malaysia. This study presents an extensive comparative assessment of these two procedures in estimating different watersheds flood peak flows. A total of 22 watersheds on Langkawi Island were simulated, representing a range of catchment types and characteristics, with catchment areas ranging from 1 to 68.2 km², and slopes between 1.54% to 24.77%, encompassing small to moderately large basins with gentle to steep terrain. For HP11 analysis, a Microsoft Excel spreadsheet was developed and used to perform the analysis based on triangular distribution of direct runoff following Snyder’s method. On the other hand, HEC-HMS software was used to perform hydrological analysis based on HP27 guidelines for both current and future land-use scenarios following Clark Unit Hydrograph Method. Obtained results were analyzed statistically using SPSS covering several parameters and tests. The results showed that the estimated design peak flows using HP11 procedure were higher compared with those estimated using HP27 for same land-use condition especially at higher ARIs. The percentage difference reached more than 70% depends on catchment characteristics. The estimated design peak discharge flows for future land use distribution were increased as expected due to the increment in the imperviousness of the catchment area, which contributes to higher peak discharge values as the critical storm duration shifts to a shorter period. Overall, the study demonstrates that the choice of hydrological procedure has a significant impact on flood estimation and should be carefully considered in future planning and design.

Influence of Machining Parameters on Surface Roughness in Trochoidal Milling of Ti-6Al-4V via Response Surface Methodology

2026-03-12T13:35:02+07:00

Machining of Ti-6Al-4V titanium alloy poses significant challenges, mainly due to its low thermal conductivity and tendency to form a built-up edge (BUE), which can lead to high surface roughness (Ra) and rapid tool wear. In this context, the trochoidal milling strategy emerges as a promising solution, as it can reduce thermal and mechanical loads by using a spiral-shaped tool path with intermittent contact between the tool and Ti-6Al-4V. Therefore, this research aims to evaluate the effectiveness and optimise the machining parameters, specifically, cutting speed (v_c) and feed rate (v_f), when using the trochoidal strategy for machining Ti-6Al-4V, with the primary objective of minimising Ra. The methodology used is Response Surface Methodology (RSM) with a Central Composite Design (CCD) to model the nonlinear relationships among v_c, v_f, and Ra. The experiment was conducted using a 3-axis CNC machining centre with carbide cutting tools. Based on the optimisation results, the optimal parameter combination to minimise Ra is 155 m/min for v_c and 124 mm/min for v_f. Experimental validation tests of the predicted optimal parameters yielded an Ra value of 0.118 μm, with an error of 2.47% relative to the RSM-predicted value, confirming the model’s applicability. In conclusion, this research confirms that targeted parameter optimisation using RSM is highly effective in improving the surface quality of Ti-6Al-4V, further proving the potential of the trochoidal strategy as an efficient method for critical applications in the aerospace and biomedical sectors where surface integrity is a prerequisite.

One-dimensional Photonic Crystal with Defects to Detect Gelatine Containing Water

2026-06-18T11:32:47+07:00

Gelatine is commonly found in food, pharmaceutical, health, and cosmetic products. However, its use has generated significant debate due to concerns about the numerous sources, particularly for religious reasons. For Muslims and Jews, gelatine derived from porcine sources must be avoided, as it is strictly prohibited by religious laws. The detection in these products is important, and this led to the design of an optical biosensor based on a one-dimensional (1D) photonic crystal (PhC) with a defect layer. The aim was to detect the distinctive characteristics of gelatine with high sensitivity. Additionally, the defect layer was placed in the center of the structure, where the gelatine solution was introduced. The designed PhC structure consisted of N layers of (ZnO/SiO₂)ᴺ/defect/(ZnO/SiO₂)ᴺ. The reflectance and transmittance characteristics of the PhC were calculated by using the transfer matrix method (TMM), performed with the MATLAB software. For simplicity, the normal incidence of light on the PhC was assumed to avoid the complexities of oblique angles. Sensor performance was evaluated by calculating sensitivity (S), quality factor (Q), figure of merit (FOM), resolution (RS), and detection limit (LOD). The biosensor achieved an average S, Q, FOM, RS, and LOD of 225.7694 nm/RIU, 2249.413, 806.3185, 0.1332, and 4.13 × 10⁻⁵, respectively. These metrics showed the simple PhC design was a highly sensitive, non-invasive detection platform for monitoring and identifying gelatine-based products that did not meet Halal and Kosher requirements. The novelty of the study centered on the application of a simple 1D PhC with a defect cavity to detect gelatine concentrations in both aqueous and oily solutions, a problem that had not been addressed in detail.

Valorization of Banana Peel Extract for Green Synthesis of Silver Nanoparticles: Physicochemical Characterization and Antibacterial Performance

2026-07-02T11:44:55+07:00

The growing demand for sustainable nanomaterial synthesis has driven interest in plant-mediated approaches that align with circular economy principles. This study presents a facile approach for generating silver nanoparticles (AgNPs) using Saba banana peel extract (Musa acuminata × balbisiana) as reducing and stabilizing agents. AgNPs were synthesised by reacting different ratios (1:1–1:12) of banana peel extract (BPE) with silver nitrate (AgNO₃, 1 mM) at 90°C for 2 h. Based on the UV-Visible spectra, the optimal BPE:AgNO₃ ratio was identified at 1:11. Fourier-transform infrared (FTIR) spectroscopy revealed that functional groups in the extract including hydroxyl, carbonyl and amine, were involved in the reduction of silver ions. Morphological analyses using Scanning electron microscopy (SEM) and Transmission Electron Microscopy (TEM) show the formation of spherical AgNPs with average size of 5.37 ± 1.36 nm, while energy-dispersive X-ray (EDX) analysis indicated a high silver content (73.87 wt%). The biosynthesised AgNPs displayed potent antibacterial activity against Staphylococcus aureus and Escherichia coli. These findings highlight the potential of Saba banana peel, an abundant and underutilized biomass, as a low-cost and sustainable resource for the green synthesis of AgNPs with promising antimicrobial properties and potential applications in biomedical, environmental, food and energy sectors.

Hybrid Nanofluid-MQL and Cold Air Cooling for Hard Milling: RSM–PSO Multi-response Optimization

2026-05-22T12:00:47+07:00

This study investigates a hybrid cooling–lubrication approach combining nanofluid-assisted minimum quantity lubrication (MQL) with cold-air cooling in the hard milling of SKD11 steel, aiming to improve both surface quality and productivity. Response surface methodology (RSM) was employed to model the relationships between machining parameters and two key performance indicators, namely surface roughness (Ra) and material removal rate (MRR). The developed models were integrated with particle swarm optimization (PSO) for multi-objective optimization. Compared with conventional MQL, the hybrid approach reduced Ra by approximately 5–10% under similar cutting conditions, indicating enhanced cooling and lubrication performance. The optimization results revealed distinct trade-offs between surface quality and productivity. When surface quality was prioritized (wRa = 0.7), a minimum Ra of 0.160 µm was achieved with a relatively low MRR of about 653 mm³/min. In contrast, both the balanced (wRa = 0.5) and productivity-oriented (wRa = 0.3) scenarios converged to the same optimal solution (Ra ≈ 0.245 µm, MRR ≈ 1790 mm³/min). This convergence suggests that MRR dominates the optimization when its weighting is comparable to or higher than that of Ra, indicating a plateau region on the trade-off surface. The proposed RSM–PSO framework provides both an effective optimization approach and new insight into balancing surface integrity and productivity in hybrid-cooled hard milling

Probabilistic vs Deterministic Foundation Design: Impact of PDA and SLT Data on Required Safety Factor in an Indonesian Project Case Study

2026-03-10T09:03:35+07:00

Foundation design using the Allowable Stress Design (ASD) method is characterized by less consistent reliability that is difficult to quantify, leading to overly conservative or optimistic design. Therefore, this study aimed to implement the Reliability-Based Design (RBD) method in foundation design to achieve consistent reliability levels and evaluate the equivalent safety factor generated. The design process was initially conducted based on API method and updated using field test data, namely the Pile Driving Analyzer (PDA) and Static Load Test (SLT). Statistical parameter updating was performed using a statistical analysis through the Bayesian updating method. The results showed that RBD method was capable of explicitly quantifying uncertainties. Furthermore, the availability of more and higher-quality data contributed to a more accurate representation of uncertainty, causing a greater reduction in the required equivalent factor of safety. This showed that the probabilistic method enabled the application of a lower safety factor compared to the conservative value of approximately 2.5 commonly adopted in ASD, without compromising the target reliability. Through the addition of data from PDA and SLT testing, this study showed that the RBD method could significantly reduce the required safety factor to a value of 1.68.

Analysis of the Effect of Permanent Magnet Width on Cogging Torque of Interior LSPMSM

2025-11-10T09:09:46+07:00

The line-start synchronous permanent magnet synchronous motors have the advantages of high efficiency and line-starting capability. However, this type of motor has the drawback of high cogging torque, which is the main cause of vibration and noise during operation. Analyzing the factors affecting cogging torque is essential and important in the design of these motors. This paper studies the impact of magnet width on cogging torque. Theoretical analysis and finite element method simulations for the test motor show that the cogging torque can be significantly high, reaching up to 5% of the rated torque of the motor. Research results indicate that cogging torque can be reduced by decreasing the width of the permanent magnet. However, reducing the magnet width adversely affects the startup characteristics, torque, and efficiency of the motor, demonstrating the existence of an optimal width that balances cogging torque and other parameters in motor design. Additionally, the research results have been validated through experimental modeling, showing a close correlation in the shape of the cogging torque and the back eletromotive force characteristics, with a small error of less than 5%.

Mitigating Maloperations of Generator Protection in Oil and Gas Facilities: A Coordination Strategy for Overcurrent and Overcurrent Voltage Restrained Relays

2025-12-02T09:51:56+07:00

The oil and gas industry requires a highly reliable and adaptive electrical protection system to prevent maloperation and ensure operational continuity. Unlike conventional industrial systems, electrical configurations in oil and gas facilities often feature short electrical distances between generators and loads, with some connected directly and others through transformers. This condition increases the risk of miscoordination in overcurrent relays (50/51 and 50/51V), potentially leading to widespread blackouts that may disrupt production processes and cause significant financial losses. Such maloperations are often linked to improper relay settings. Inaccurate modelling of time–current characteristic (TCC) curves and misadjustment of 50/51V relays can result in non-selective tripping, where the 50/51V relay operates before the 50/51 relay as the primary protection. This study proposes relay coordination strategies for three representative power system configurations in the oil and gas sector, which are evaluated through relay coordination simulations using power system simulation software. The approach considers for the unique characteristics of the 50/51V relay and equipment damage curves, using power system simulation software for evaluation. The results demonstrate that the proposed scheme avoids maloperation by maintaining IEEE standard coordination margins (0.2–1 s), ensuring relay curves remain below electrical equipment damage thresholds, preserving selectivity and preventing relay coordination maloperations. These findings provide practical guidelines for engineers in designing selective and reliable protection systems, thereby reducing blackout risks and supporting continuous operation in oil and gas facilities.

Defect-Engineered TiO2 Photocatalysts for Enhanced Methylene Blue Degradation

2026-07-03T19:11:31+07:00

Textile wastewater containing synthetic dyes is a significant environmental concern due to its high color intensity, toxicity, chemical stability, and resistance to conventional treatment. This work prepared Ti³⁺ self-doped TiO₂ photocatalysts using a simple sol-gel route combined with NaBH₄-assisted calcination. The strategy was designed to generate intrinsic lattice defects, mainly Ti³⁺ centers and oxygen vacancies, in anatase TiO₂ without the use of flammable reducing gases or complicated post-synthesis treatments. The influence of calcination temperature was examined by preparing samples at 200, 500, and 600 °C. XRD results showed that the selected samples maintained the anatase TiO₂ phase, indicating that the reduction-calcination treatment did not produce detectable secondary crystalline phases. Higher calcination temperature increased the degree of crystallinity, with crystallinity values of 59.1%, 68.2%, and 69.6% for the samples of T-200, T-500, and T-600, respectively. Raman spectra confirmed the anatase framework and revealed features related to lattice disorder in the reduced samples, while EPR analysis directly confirmed Ti³⁺/oxygen-vacancy defect sites, with the strongest signal observed for T-600. UV-vis DRS demonstrated that the samples calcined at 500 and 600 °C absorbed more strongly in the visible region. The T-600 sample showed an effective band gap of 2.42 eV, evidencing that defect-related states contributed to broader light harvesting. Photocatalytic evaluation indicated that methylene blue degradation improved with increasing calcination temperature, and the best performance was obtained using T-600, with an apparent rate constant of 0.017 min^-1. Reusability testing showed that T-600 retained activity for three cycles, although the rate constant decreased to 0.0116 and 0.0089 min^-1 in the second and third cycles, respectively. The enhanced activity is attributed to the combined contribution of higher anatase crystallinity, improved visible-light response, and an appropriate concentration of Ti³⁺-oxygen vacancy sites that support charge separation and reactive oxygen species formation.