Institutional Repository

U n i v e r s i t y of P e r a d e n i y a

The Institutional Repository of the University of Peradeniya is the University's digital gateway to scholarly knowledge and research excellence. Established as the next phase of the Digital Library initiative launched by the University Library in 2011, it preserves, showcases, and provides open access to the intellectual output of the University community.

Through a rich collection of research articles, conference proceedings, theses, dissertations, and other scholarly works, the repository enhances the visibility, accessibility, and long-term preservation of knowledge, extending the global reach and impact of the University of Peradeniya’s academic and research contributions.

Photo by @ITD,Main Library

Communities in DSpace

Select a community to browse its collections.

Now showing 1 - 5 of 5

Recent Submissions

Feasibility study and implementation of electric vehicle conversions for motorcycles in Sri Lanka
(University of Peradeniya, Sri Lanka, 2026-08-20) Jayawardhana, H. D. S. M.; Ranathunge, K. U.; Bakmeedeniya, L. U.; Wickramarathna, T.
The transition away from fossil fuel dependency in Sri Lanka’s transport sector is critical due to rising fuel costs and environmental concerns. As motorcycles dominate personal transport, this study evaluates the technical feasibility and economic viability of converting internal combustion engine (ICE) motorcycles to electric vehicles (EV). Conducted in collaboration with national energy sector experts, the primary objective was to develop a scalable, low-cost conversion model using the widely used Bajaj CT100 as a prototype. While previous local conversion studies have focused primarily on simulation or basic assembly without real-world context, this research establishes a novel, comprehensive framework by bridging the gap between rigorous technical testing, user-centric surveys, and detailed financial forecasting. The methodology integrated a market survey of over 100 riders to identify user preferences, followed by the selection of a 1000W brushless DC (BLDC) motor and a 48V controller system. A 3D model of the modified drivetrain and mounting brackets was developed using SolidWorks, and weight distribution calculations were performed to ensure vehicle stability. Physical testing was conducted in a laboratory environment using a regulated power supply to verify the responsiveness and mechanical compatibility of the conversion kit. Financial analysis revealed that the conversion offers substantial annual savings of over LKR 80,000 compared to petrol operation, with an estimated payback period of approximately two years. Technical results confirmed a safe weight distribution of 53.4% rear and 46.6% front, even with the added weight of the hub motor. Furthermore, to mitigate range anxiety, a novel plug-in hybrid configuration was conceptualized. This configuration proposes retaining the original ICE drivetrain components alongside the newly integrated rear-wheel electric hub motor, allowing the rider to manually switch between petrol and electric modes using custom switching relays and clutch-based control. The study concludes that EV conversion is a viable and cost-effective strategy for sustainable mobility in Sri Lanka. These findings provide a technical foundation for localized standardized conversion kits and highlight a significant market opportunity for cleaner transportation.
Explainable AI-driven zero-trust anomaly detection for encrypted traffic
(University of Peradeniya, Sri Lanka, 2026-08-20) Wanasinghe, J. K.; Perera, M. C. S. J.; Wijewardhana, A. R. S. S.; Karunarathna, S. N.; Jayasinghe, U.
Modern network attacks increasingly exploit encrypted channels such as TLS and HTTPS, rendering traditional deep packet inspection ineffective. Existing intrusion detection systems achieve high accuracy but incur latency unsuitable for real-time software-defined networking (SDN) deployment, and provide limited attribution for automated blocking decisions required by zero-trust architectures. Prior work has advanced encrypted-traffic detection via deep autoencoders, hybrid CNN architectures, contrastive learning, and SHAP-based classifiers, yet each trade off encrypted-traffic compatibility, latency, explainability, or real-capture validation; jointly satisfying all four remains comparatively unaddressed. This paper presents a two-stage anomaly detection pipeline built around this combination, operating exclusively on IP and TCP packet header metadata without payload decryption. The design was shaped by comparative experimentation on the BCCC Darknet (25,588 unlabelled samples) and the CIC-IDS-2017 (702,007 expert-annotated flows); only the latter, split into a stratified training set with a held-out real-world test set, trained and evaluated the final pipeline, with the Darknet experiments used solely to motivate its consensus-voting design. Stage 1 deploys a Base Check Classifier (BCC), a Decision Tree trained on 28 header-only features selected through a four-stage Gini-importance and correlation-based reduction from 79 CICFlowMeter features, achieving 99.964% attack recall at ~1.2ms per flow. Stage 2 analyses only BCC-flagged flows using Deep Dictionary Learning (DDL) and an Isolation Forest consensus voter, whose dual-model agreement reduces the false positive rate relative to either model alone. Explainability combines DDL-native per-feature reconstruction error decomposition, SHAP, and LIME, with cross-method agreement attributing every DROP decision. A novel PCAP-to-labeled-stream construction methodology maps 579,932 raw streams to expert-annotated ground truth via 5-tuple and timestamp matching, enabling direct evaluation on real packet captures (not pre-extracted features). On 3,000 such streams, the pipeline attains 96.69% accuracy, 99.31% attack recall, 91.28% precision, and a 0.69% false negative rate, matching or exceeding deep-learning and SHAP baselines on the same benchmark. Physical validation on a 3-computer inline bridge testbed enforcing iptables DROP-by-default confirms real-time zero-trust enforcement with Grafana monitoring. These results indicate that header-only two-stage detection with multi-method explainability is a measurably effective approach for zero-trust encrypted-traffic anomaly detection; generalisation to newer datasets and live traffic is the next validation step.
Experimental investigation of local heat transfer coefficients of a minichannel evaporator with propane as the refrigerant
(University of Peradeniya, Sri Lanka, 2026-08-20) Malshani, H. A.; Fernando, W. P. D.
Accurate heat transfer coefficients are fundamental to the design and optimization of refrigeration systems. This study investigates the local heat transfer coefficients of a minichannel heat exchanger used as an evaporator in a heat pump with propane (R290) as the refrigerant. The growing demand for sustainable refrigeration systems is driving the adoption of natural refrigerants with negligible ozone depletion and low global warming potential. Propane offers excellent thermodynamic performance, but its flammability requires charge minimization, highlighting the importance of compact heat exchangers with small internal volumes and high efficiency. The experimental data were obtained from a previously developed minichannel aluminum evaporator operating in a controlled system. Previous studies by one of the authors focused on determining average heat transfer coefficients of the evaporator. In contrast, the present study utilizes the same experimental dataset to investigate the local heat transfer characteristics of propane in the minichannel evaporator through a detailed segment-wise analysis of heat transfer behavior. The evaporator was divided into 12 sections to calculate local shell-side heat transfer coefficients. To evaluate the applicability of existing predictive models, the experimentally obtained heat transfer coefficients were compared with several widely used empirical and semi-empirical correlations from the literature. Where necessary, constants were adjusted while preserving the original mathematical formulation of each correlation. These included correlations proposed by Lazarek and Black, Malek and Colin, Pierre, Cooper, Kandlikar, which have been extensively applied in the analysis of boiling heat transfer in conventional and microscale channels. Among the evaluated correlations, the Kandlikar model exhibited the best agreement with experimental data across all flow regimes, demonstrating adaptability to minichannel conditions. The Malek and Colin correlation performed well in the saturated boiling region after modification, indicating its suitability for predicting nucleate flow boiling under certain conditions. In contrast, Lazarek and Black and Pierre correlations showed moderate agreement, with limitations in accurately capturing superheated or transitional regions. These findings demonstrate that section-wise analysis provides a more accurate representation of boiling heat transfer than average-value methods and offers improved predictive capability for the design and optimization of compact, low-charge propane heat exchangers used in sustainable refrigeration and heat pump applications.
Evaluating out-of-domain performance prediction methods under domain shifts
(University of Peradeniya, Sri Lanka, 2026-08-20) Weerasinghe, P.; Tennakoon, R.; Thangarajah, J.; Herath, D.
Deep neural networks have achieved remarkable success in various applications where the test data follows the same distribution as the training data. However, in many real-world scenarios, this assumption often fails due to domain shifts, where test data referred to as out-of-domain (OoD) data differs from the training distribution. Such shifts can significantly degrade model performance and are particularly pronounced in data-constrained domains such as medical imaging, arising from variations in imaging devices, acquisition protocols, and patient demographics. While numerous techniques have been proposed to address domain shift, their effectiveness remains limited, especially when labeled target data is unavailable. This study investigates algorithms designed to predict OoD performance on unlabeled datasets and evaluates their reliability under different types of domain shifts, specifically diversity shifts (emergence of new patterns) and correlation shifts (changes in feature–label relationships). A systematic benchmarking framework is adopted to evaluate a diverse set of state-of-the-art OoD performance prediction algorithms across datasets exhibiting both shift types. The evaluation includes benchmark datasets such as PACS, CIFAR-100- C, and ColoredMNIST. OoD performance prediction methods Average Threshold Confidence (ATC), Difference of Confidence (DoC), Mandoline, and AugMax are assessed under consistent experimental settings using metrics such as robustness accuracy and mean absolute error. The results reveal significant variability in algorithm performance across shift types, indicating a lack of consistent generalization. Under correlation shift (MNIST to ColoredMNIST), ATC achieves substantially higher robustness accuracy (0.9466) compared to DoC (0.2096), demonstrating its effectiveness in confidence-based estimation. In contrast, under diversity shift (CIFAR-100 to CIFAR-100-C), Mandoline outperforms ATC (0.5164 vs. 0.2655), highlighting its sensitivity to distributional variation. Overall, no single method performs reliably across both shift types, emphasizing limitations of current approaches in complex real-world scenarios. These findings highlight the need for shift-aware or hybrid OoD performance prediction frameworks capable of handling diverse and interacting domain shifts. This work provides a benchmark for future research and offers critical insights for advancing reliable OoD performance estimation.
Enhancing developer productivity through hybrid knowledge graph–LLM integration in software engineering environments
(University of Peradeniya, Sri Lanka, 2026-08-20) Vaishaly, M.; Vasanthapriyan, S.; Rubasinghe, T. D.
Modern software repositories contain highly interconnected artifacts, including source code, documentation, issue reports, pull requests, commits, and dependency information, making repository exploration and software maintenance increasingly challenging. Although Large Language Models (LLMs) provide valuable assistance for software engineering tasks, they frequently lack repository-specific contextual understanding and may generate unsupported or hallucinated responses when reasoning across multiple software artifacts. This study proposes an adaptive hybrid Graph Retrieval-Augmented Generation (GraphRAG) framework that integrates a repository-specific knowledge graph with graph and vector retrieval to generate accurate, explainable, and evidence-grounded repository reasoning. The research addresses the questions of whether graph-grounded retrieval improves repository question answering, reduces hallucinations, supports multi-hop reasoning, and enhances repository exploration efficiency. A mixed-method evaluation is adopted using both a large-scale synthetic software engineering dataset and five real-world open-source repositories: Spring PetClinic, OpenMRS, eShopOnContainers, Apache Airflow, and MLflow. The synthetic dataset is programmatically generated to simulate realistic software development environments by creating interconnected repositories, developers, commits, issues, pull requests, documentation, dependencies, and software components with predefined semantic relationships. This controlled environment provides complete ground truth, enables reproducible experiments, supports scalability analysis, and facilitates systematic ablation studies, while the real repositories validate the framework under practical software engineering conditions. Repository artifacts are parsed using static code analysis and repository metadata extraction, transformed into a Neo4j Knowledge Graph, and integrated with adaptive GraphRAG retrieval. A benchmark comprising 300 manually designed repository reasoning questions is developed across entity retrieval, dependency analysis, issue reasoning, documentation retrieval, pull request analysis, and multi-hop reasoning. Ground-truth answers are generated through validated Cypher queries and manual verification against repository evidence. Experimental evaluation against Plain LLM, Vector RAG, Graph RAG, and the proposed adaptive GraphRAG framework demonstrates improvements with an Exact Match score of 0.842, Precision of 0.891, Recall of 0.876, F1-score of 0.883, ROUGEL of 0.901, BERTScore of 0.928, a hallucination rate of 0.071, multi-hop reasoning accuracy of 0.864, and an average repository task completion time of 41.8 seconds, outperforming the comparison methods. These findings demonstrate that adaptive graph-grounded retrieval provides more reliable, explainable, and scalable repository reasoning while establishing a reproducible evaluation framework for intelligent software engineering assistants.