Thirumalai Engineering College

The operational durability of Wireless Sensor Networks (WSNs) used in environmental monitoring, healthcare and industrial automation decreases due to sensor node failures, energy depletion, and communication breakdowns. The fault tolerance and the optimization of energy used in the operations are necessitated by the issues that develop as a result of these problems. The existing fault-tolerant methods do not provide an efficient fault recovery with low latency and high network performance. The study presents the concept of BIAFTEL (Bio-Inspired AI-Driven Frameworks of Fault Tolerance and Extended Lifespan) that unites bio-inspired learning algorithms with the predictive fault tracking of AI-based methods to augment the fault tolerance and long-life time of WSNs. The biological resilience model affects BIAFTEL because of its applications of self-healing techniques and swarm intelligence combined with adaptive behaviours, which allow automated fault-detection, routing optimization, and network outage. The main innovation of the system identifies faults through LSTM Autoencoders based fault detection techniques which train normal operating patterns to spot anomalies before they happen hence enhancing prediction precursors. HABCO serves as the routing mechanism because it enables adaptive energy-efficient networking that allows for lower power usage and better network lifetime extension. The network operation is executed through Python programming. Experimental results show that BIAFTEL achieves a 99.3

Sub-6 GHz $\mathbf{5 G}$ wireless systems require antennas that are compact, economical, and efficient, while remaining easy to integrate into user equipment and small-cell platforms. Dualband microstrip patch radiators are well suited for this role, yet their design must balance footprint, impedance bandwidth, and radiation characteristics, particularly when realized on lossy substrates such as FR-4. Many existing designs rely on manual parameter sweeps or conventional metaheuristic algorithms, which often converge slowly, favour larger electrical dimensions, and provide only moderate bandwidth, especially when dualband performance is considered. In addition, several reported optimization studies concentrate on a single operating band or do not explicitly improve both impedance matching and gain at the two resonant frequencies. This work introduces an Improved Harris Hawks Optimization (HHO) scheme for the synthesis of a dual-band microstrip patch antenna designed to operate at 3.50 GHz and 4.90 GHz on a $30 \times 30 ~\text{mm}^{2}$ FR-4 substrate. The optimizer adjusts six key geometrical variables associated with the patch, slots, and parasitic stubs using a dual-frequency fitness function that minimizes the combined reflection coefficient at the two target bands. A large set of candidate geometries is evaluated through full-wave electromagnetic simulations across the sub- 6 GHz region, which effectively generates a design dataset capturing the nonlinear relationship between structural parameters and antenna responses. This dataset serves to benchmark the Improved HHO against recent nature-inspired optimizers, including SMA, AO, MPA, and GBO, under identical stopping conditions. The optimized antenna achieves $S_{11}$ levels below $-24 \mathbf{~ d B}$, with $-10 \mathbf{~ d B}$ bandwidths of $\mathbf{1 2. 0 \%}$ and 7.3%, realized gains of 5.12 dBi and 6.05 dBi, and radiation efficiencies above 83 % in both bands. These results indicate that the Improved HHO framework yields a compact, high-performance dual-band antenna suitable for 5 G applications and can be adapted to the design of more complex multi-band or array configurations.

Heat sinks are critical components for absorbing and dissipating heat in Variable Refrigerant Flow (VRF) systems, with performance largely determined by material properties and surface characteristics. Recent studies have demonstrated that nano-based coatings can significantly enhance heat transfer, with titanium dioxide ($\text{TiO}_{2}$) nanoparticles offering favorable thermal and mechanical properties for improved dissipation. In this study, $T i O_{2}$ nanoparticles were applied to heat sink surfaces, and their effects on thermal performance were systematically investigated. Key input parameters included coating thickness, nanoparticle concentration, and airflow rate, while the primary responses were heat flux, temperature drop, and coefficient of performance (COP). Using Response Surface Methodology (RSM) with Central Composite Design (CCD), the experiments and optimization revealed that nanoparticle concentration and airflow rate were the most influential factors, with coating thickness playing a moderate role. The optimized conditions- $0.496 \text{wt} \%$ nanoparticle loading, 415 nm coating thickness, and $254 \mathrm{m}^{3} / \mathrm{h} / \text{kw}$ airflow rate-led to significant enhancements in thermal performance. These findings demonstrate that parameter optimization through RSM can effectively improve the efficiency and reliability of VRF system heat sinks, providing practical guidance for advanced thermal management applications.

To optimize the weights and biases of the LSTM, this work introduces a novel hybrid model called TLBO-LSTM, which is based on an evolutionary algorithm (EA) TLBO and uses a less parametric algorithm. After that, we use this hybrid model to look at how well four major stock indexes have been predicting their closing prices. Two more models, GWO-LSTM and PSO-LSTM, were built using GWO and PSO, respectively, and are also included in the forecasting assignment; their purpose is to compare the performance of the proposed model. Using MAPE and RMSE error measures, we evaluate each model’s prediction abilities. The experimental findings show that TLBO-LSTM is the best model compared to its competitors.

This study investigates the copyrolysis of waste paint oil (WPO) and algae biomass as an integrated waste-to-energy pathway for producing high-energy biochar while mitigating the environmental risks associated with WPO disposal. Copyrolysis experiments were conducted in a laboratory-scale fixed-bed reactor using WPO–algae weight ratios of 1 : 0.5 to 1 : 1.5 under an inert nitrogen atmosphere (150 mL/min), with temperatures ranging from 400 to 600°C, a heating rate of 10°C/min, and a residence time of 60 min. The physicochemical properties of WPO (density 0.88–0.89 g/mL; kinematic viscosity 4.5–5.0 mm2/s) were controlled to ensure feedstock consistency. The resulting biochars exhibited markedly improved solid-fuel characteristics compared to single-feed pyrolysis. Fixed carbon content increased from 65.3 to 72.8 wt