Government Engineering College, Jehanabad

In this work, the application of machine learning (ML) models for the prediction of titanium alloy for electrical discharge machining (EDM) was studied. The machining performances (MRR, TWR, Ra , and recast layer thickness) were predicted by SVM, XG Boost and KNN machine learning models. The heatmap highlighted a strong correlation between parameters and a positive effect was observed for MRR (0.96) and Ra-recast layer thickness (0.82). The performance metrics such as Mean Absolute Error (MAE), Mean Squared Error (MSE), and Root Mean Squared Error (RMSE) were also utilized to compare the models. SVM achieved the best prediction results for TWR and recast layer thickness, while KNN produced the lowest MAE on MRR. The XG Boost model had slightly compromised prediction accuracy than the support vector machine for TWR. These results suggested that a hybridized approach could improve prediction accuracy for complex machining parameters, especially in high-precision machining applications such as biomedical and aerospace industries.

The rapid evolution of industrial automation has increased the demand for reliable wireless communication and embedded control integration. Conventional System-on-Chip architectures have lacked the capability to jointly optimize microwave behavior and embedded processing, which has resulted in performance degradation under industrial interference conditions. This study has addressed the challenge by proposing a Microwave-Aware Circuit-Embedded Co-Design (MACE-CoD) method, that has integrated RF circuit characteristics with digital control subsystems at the design stage. The proposed method has considered the electromagnetic interactions, that which has influenced signal integrity, latency, and power consumption across multi-domain components. A co-simulation framework has been developed, which has combined microwave circuit modeling with embedded firmware optimization. The architecture has enabled adaptive tuning mechanisms, that which has responded to environmental variations in real time. Experimental results show that the proposed MACE-CoD method achieves an SNR of 26 dB, that which improves performance by 44%, reduces latency to 5 ms, that which decreases delay by 37%, and lowers energy consumption to 42 mW, that which enhances efficiency by 30%. The system attains a packet delivery ratio of 98% and reduces the bit error rate to 0.01, that which ensures high reliability under industrial interference conditions.

This study proposed a hybrid machine learning framework integrated with reliability analysis to predict the factor of safety against liquefaction and quantify the associated uncertainties. Three optimized Extreme Gradient Boosting models, XGB-AOA, XGB-ESMA, and XGB-MPA, were developed and evaluated using comprehensive statistical, graphical, and probabilistic analyses. Model performance was assessed through actual and predicted comparisons, multi-criteria score evaluation, Regression Error Characteristic curves, Anderson-Darling testing, internal k-fold cross-validation, and external validation. Reliability-based methods, including the first-order reliability method, second-order reliability method, and higher-order approximations, were applied to estimate the probability of failure and reliability index. Model interpretability was investigated using SHapley Additive explanations (SHAP) to identify the most influential variables governing liquefaction resistance. Results indicated that the XGB-ESMA model achieved superior prediction accuracy, robustness, and generalisation capability compared with XGB-MPA and XGB-AOA. Reliability analysis revealed a high probability of failure (approximately 60–64

Predictive maintenance has become an important component of intelligent industrial systems by enabling machinery condition assessment before critical failures occur. This study developed a data-driven digital-twin framework integrating machine learning with temporal sensor analytics for industrial machinery health monitoring. The analysis used 1,000 chronologically ordered observations comprising vibration, temperature, pressure, and machinery fault states. The dataset included normal operation (60.9%), bearing faults (30.3%), and overheating (8.8%). Sensor measurements were standardized and transformed using 128-observation sliding windows, producing 44 multivariate temporal samples across three sensor channels. Shapelet-based temporal feature extraction was incorporated to capture condition-related patterns, while Logistic Regression and Random Forest were considered for machine-state classification. The results demonstrated substantial variability in the dynamic sensor measurements but weak individual relationships with machinery fault status. Vibration exhibited the highest absolute sensor–fault correlation (r = −0.032), followed by pressure (r = 0.009) and temperature (r = −0.008), indicating that individual sensor measurements alone provided limited linear discrimination between operating states. The proposed framework therefore emphasizes multivariate temporal behavior for updating digital-twin machinery states and supporting condition-based maintenance decisions. The findings demonstrate the feasibility of combining temporal sensor processing, machine learning, and data-driven digital twins for industrial condition monitoring, while further validation using larger run-to-failure datasets is required.

This study developed an integrated machine learning framework for predicting the factor of safety (FOS) of soil slopes using Extreme Gradient Boosting (XGB) optimized with three metaheuristic algorithms: Particle Swarm Optimization (PSO), Tiki Taka Optimization Algorithm (TTOA), and Whale Optimization Algorithm (WOA). A dataset comprising six influential parameters, namely cohesion, unit weight, friction angle, slope angle, slope height, and pore pressure ratio, was employed for model development. The dataset was normalized and divided into training and independent testing subsets, while five-fold cross validation was adopted during model optimization to improve robustness and reduce overfitting. Model performance was assessed using multiple statistical indicators, scatter plots, Regression Error Characteristic curves, and Taylor diagrams. The results demonstrate that all optimized XGB models effectively captured the nonlinear relationship between the input parameters and FOS, while the proposed framework provides an accurate, interpretable, and computationally efficient tool. To improve model transparency, SHAP, LIME, and permutation feature importance analyses were incorporated. The explainability results identified slope angle, cohesion, and slope height as the dominant predictors, consistent with their governing roles in slope geometry and shear strength. The proposed framework provides an accurate, interpretable, and computationally efficient tool for data driven slope stability assessment.