Kathmandu University

This study produces a multi-year flood susceptibility assessment for the Koshi River Basin, Nepal, by integrating Sentinel-1 SAR imagery with machine-learning and geographic information systems (GIS)-based approaches. Five-year (2019-2023) temporal flood analysis from Sentinel-1 GRD data was used for the preparation of flood inventory, consisting of 375 flood and 375 non-flood points by extracting multi-year composites alongside 11 environmental, topographic, climatic, and vegetation-based conditioning factors, which served as the foundation for training and validating the susceptibility models. The four machine learning (ML) classifiers: bagging, random forest (RF), support vector machine (SVM), and AdaBoost were evaluated against the weighted sum method. Bagging achieved the highest performance (accuracy 94.52%, precision 95.45%, recall 95.45%, F1-score 95.45%, and AUC 0.990) and identified 6.185% of the basin as highly susceptible to flooding. Strong agreement between bagging and RF (r = 0.76) indicated consistent model behaviour, while weaker correlations with the weighted sum method underscored limitations of rule-based approaches in capturing nonlinear flood processes. The combination of a multi-year SAR-based flood inventory and ensemble ML techniques demonstrates clear advantages in generating reliable susceptibility maps, providing actionable evidence to support land-use planning, disaster preparedness, and targeted flood mitigation.

Greenhouse gas emissions from sanitation systems remain underquantified, particularly when considering the entire service chain. Previous studies have largely focused on emissions from containment, with limited attention to later stages such as collection, transport, treatment and disposal. To address this gap, this research comprehensively estimates greenhouse gas (GHG) emissions from sanitation systems in Lahan municipality, Nepal. We used an extended version of the IPCC-based Tier-1 approach. Data collection included a household survey and key informant interviews. In scenario A, the baseline total annual emissions are 8.7 Gg CO2e, mostly from the digestion of faecal sludge in the containment (7.3 Gg CO2e). In scenario B, when a projected faecal sludge treatment plant (FSTP) is built and in operation, annual emissions reach 10.0 Gg CO2e, driven by methane emitted by the anaerobic digester in the plant. Scenario C considers climate mitigation strategies: increasing the share of households emptying their containments, increased emptying frequency and adding of methane capture in the FSTP. This can reduce annual emissions to 7.9 Gg CO2e per year, which is 21% less than in scenario B. Our results suggest that methane capture in the FSTP is the most critical mitigation strategy.

The electric distribution system is undergoing rapid integration of new technologies such as renewable energy resources, energy storage system, electric vehicles etc. The addition of these resources can cause over-voltage, reduced line hosting capacity, and overloading of distribution assets thus requiring dynamic visualization tools to operate them in a safe and secure manner. In addition, a real-world test-bed to study the problems in electrical distribution system and algorithms for their mitigation in a controlled and secure manner is also highly needed. Taking this problems into consideration, this study presents a test-bed primarily to test the algorithms for dynamic and predictive visualization, and electrical fault localization. The test-bed is a real-world operational distribution system of Kathmandu University which constitutes a supply transformer, electrical distribution lines (overhead and underground), and electrical loads. A total of six smart meters have been installed in the test-bed which sends data via wireless fidelity. These data are used to develop algorithms to determine and visualize the dynamic status of the test-bed. The algorithm is studied under various operating scenarios, including base, peak, and off-peak load conditions, and during fault events. The results show that the test-bed minimum voltage is 0.95 pu and the line maximum loading is 60.2

Purpose The purpose of this paper is to investigate the mediating role of satisfaction (SAT) in relation to mobile banking service quality (MB-SQ) and continuance intention (CI) among Nepali mobile banking users. Design/methodology/approach The paper adopted a quantitative approach and cross-sectional survey research design. Data were collected with structured questionnaires from 326 mobile banking users. A partial least squares structural equation modeling (PLS-SEM) and artificial neuro network (ANN) approach were applied to examine hypotheses. Findings Results confirm a significant positive influence of MB-SQ on SAT and CI of mobile banking adoption. Moreover, MB-SQ partially mediates the relationship between SAT and CI of mobile banking adoption. Research limitations/implications Based on the findings of this research, theoretically, this paper attempted to investigate the mediating role of MB-SQ in the CI of mobile banking, and managerially, mobile banking service providers could have insights on designing mobile banking service marketing strategy. Originality/value This paper is among the earliest studies to investigate the role of MB-SQ as a higher-order reflective-reflective construct on CI. Moreover, the endogeneity issue has been tested, and ANN has been applied to investigate the predictive relevance of SAT and MB-SQ on CI of mobile banking users. Furthermore, the authors have delved into the ongoing discourse surrounding Generation Y and Generation Z, exploring their implications on CI within the realm of mobile service quality. It provides a critical juncture for understanding continuance intention in the mobile service quality context.

Fault detection in power transmission lines is vital for maintaining the electrical grid stability and reliability. Conventional methods for identifying and classifying faults generally suffer from inefficiencies, inaccuracies, and poor interpretability. This study introduces an Explainable Artificial Intelligence (XAI)-driven approach to enhance fault identification and classification in transmission lines. The study uses MATLAB Simulink simulated data to evaluate and compare the performance of various Machine Learning (ML) models, including eXtreme Gradient Boosting (XGBoost), Extreme Learning Machine (ELM), Long-Short Term Memory (LSTM), Long Short-Term Memory-Gated Recurrent Unit (LSTM-GRU), and Vision Transformer (ViT) against the proposed CNN-Transformer hybrid model. Results demonstrate that the CNN-Transformer hybrid model outperforms alternative models, achieving superior fault detection accuracy across diverse resistance values. The CNN-Transformer model achieved the highest accuracy (0.97), followed by XGBoost (0.93), LSTM-GRU (0.92), LSTM (0.88), ViT (0.93), and ELM (0.87). Additionally, we implement SHAP, an explainable AI framework, to enhance transparency and interpretability in fault diagnosis. The developed methodology supports a mobile application enabling utility operators to monitor and identify faults in real time. This work illustrates ML along with XAI's potential in power system fault detection, paving the way for intelligent and reliable grid management solutions.