Faculty of Engineering
Permanent URI for this communityhttps://rda.sliit.lk/handle/123456789/4203
Browse
32 results
Search Results
Publication Open Access QPred: A Lightweight Deep Learning-Based Web Pipeline for Accessible and Scalable Streamflow Forecasting(Tech Science Press, 2026) Makumbura, R.K; Wijesundara, H; Sajindra, H; Rathnayake, U; Kumar, V; Duraibabu, D; Sen, SAccurate streamflow prediction is essential for flood warning, reservoir operation, irrigation scheduling, hydropower planning, and sustainable water management, yet remains challenging due to the complexity of hydrological processes. Although data-driven models often outperform conventional physics-based hydrological modelling approaches, their real-world deployment is limited by cost, infrastructure demands, and the interdisciplinary expertise required. To bridge this gap, this study developed QPred, a regional, lightweight, cost-effective, web-delivered application for daily streamflow forecasting. The study executed an end-to-end workflow, from field data acquisition to accessible web-based deployment for on-demand forecasting. High-resolution rainfall data were recorded with tipping-bucket gauges and loggers, while river water depth in the Aglar and Paligaad watersheds was converted to discharge using site-specific rating curves, resulting in a daily dataset of precipitation, river water level and discharge. Four DL architectures were trained, including vanilla Long Short-Term Memory (LSTM), stacked LSTM, bidirectional LSTM, and Gated Recurrent Unit (GRU), and evaluated using Nash-Sutcliffe Efficiency (NSE), Coefficient of Determination (R2), Root-Mean-Square-Error-Standard-Deviation Ratio (RSR), and Percentage Bias (PBIAS) metrics. Performance was watershed-specific, as the vanilla LSTM demonstrated the best generalisation for the Aglar watershed (R2 = 0.88, NSE = 0.82, RMSE = 0.12 during validation), while the GRU achieved the highest validation accuracy in Paligaad (R2 = 0.88, NSE = 0.88, RMSE = 0.49). All models achieved satisfactory to excellent performance during calibration (R2 > 0.91, NSE > 0.91 for both watersheds), demonstrating strong capability to capture streamflow dynamics. The highest performing models were selected and embedded into the QPred application. QPred was developed as a lightweight web pipeline, utilising Google Colab as the primary execution environment, Flask as the backend inference framework, Google Drive for artefact storage, and Ngrok for secure HTTPS tunnelling. A user-friendly front end utilises range sliders (bounded by observed minima and maxima) to gather inputs and provides discharge data along with metadata, thereby enhancing transparency. This work demonstrates that accurate, context-aware deep learning models can be delivered through low-cost, web-based platforms, providing a reproducible and scalable pipeline for hydrological applications in other watersheds and for practitioners. CopyrightPublication Embargo Enhancing the effectiveness of satellite precipitation products with topographic and seasonal bias correction(Elsevier B.V., 2026-02) Wanniarachchi, S; Sarukkalige, R; Hapuarachchi, H.A. P; Gomes, P.I.A; Rathnayake, UEstimating precipitation distribution across large regions is crucial for understanding water availability, planning infrastructure, and forecasting flood hazards. Traditional gauge-based methods face challenges, particularly with sparse gauge networks. In response, satellite-based, near-real-time (NRT) precipitation data has gained popularity, especially in poorly gauged watersheds. However, satellite precipitation data quality is often compromised by latency, atmospheric complexities, and topographic effects, resulting in nonlinear errors. To overcome the research gap, this study introduces the Heavy Rain Peak Adjustment (HRPA) method alongside the well-established Seasonal Autoregressive Integrated Moving Average (SARIMA) model for satellite precipitation bias correction. The analysis utilised Global Satellite Mapping of Precipitation (GSMaP-NRT) data and hourly precipitation records from 31 rain gauges in the Ovens River region of Australia. On average, the mean residual of observed and GSMaP-NRT precipitation was −0.02 mm. Additionally, the HRPA method yielded better linear regression R2(0.911), NSE (log) (−0.847), and RMSE (0.628) compared to SARIMA. The results indicate that HRPA outperforms SARIMA, particularly at lower elevations, whereas SARIMA struggles at higher elevations, underscoring its limitations in those areas. Additionally, autocorrelation and partial autocorrelation plots for some stations in hilly areas show significant wave-like patterns, indicating greater uncertainty in satellite precipitation estimates over complex terrain. For several stations, autocorrelations at 24 and 48-hour lags suggest a systematic influence of past residuals on future ones, emphasizing the need for further refinement in satellite precipitation correction methods for these regions.Publication Open Access Reviving Urban Landscapes: Harnessing Pervious Concrete Pavements with Recycled Materials for Sustainable Stormwater Management(Multidisciplinary Digital Publishing Institute (MDPI), 2025-10-29) Gunathilake, T.A; Siriwardhana,K.D; Miguntanna,N; Miguntanna, Nadeeka; Rathnayake, U; Muttil, NThis study examines the effectiveness of pervious concrete pavements as a sustainable and cost-effective stormwater management technique, particularly by incorporating locally sourced recycled materials into their design. It evaluates the stormwater treatment potential of three pervious concrete pavement types incorporating recycled plastic, glass, and crushed concrete aggregates, with six design variations produced using 25% and 50% replacements of coarse aggregates from these materials. The key properties of pervious concrete, namely compressive strength, porosity, unit weight, and infiltration, and key water quality indicators, namely pH, electrical conductivity (EC), total suspended solids (TSS), colour, turbidity, chemical oxygen demand (COD), nitrate (NO3−), and orthophosphate (PO43−), were analysed. Results indicated an overall improvement in the quality of the stormwater runoff passed through all pervious concrete pavements irrespective of composition. Notable reductions in turbidity, TSS, colour, COD, PO43−, and NO3− underscored the effectiveness of pervious concrete containing waste materials in the treatment of stormwater runoff. Pervious concrete pavements with 25% recycled concrete exhibited optimal performance in reducing TSS, COD, and PO43− levels, while the 50% recycled concrete variant excelled in diminishing turbidity. However, the study found that the use of recycled materials in pervious concrete pavements affects properties like compressive strength and infiltration rate differently. While incorporating 25% and 50% recycled concrete aggregates did not significantly reduce compressive strength, the effectiveness of stormwater treatment varied based on the mix design and type of recycled material used. Thus, this study highlights the potential of utilizing recycled waste materials in pervious concrete pavements for sustainable stormwater management.Publication Open Access Risk Evaluation of Cost Overruns (COs) in Public Sector Construction Projects: A Fuzzy Synthetic Evaluation(MDPI, 2023-04-22) Chadee, A.A; Martin, H.H; Gallage, S; Banerjee, K.S; Roopan, R; Rathnayake, U; Ray, IIn the Small Island Developing States (SIDS), public sector infrastructure projects (PSIPs) fail to both meet targeted performance metrics and deliver on the intended benefits to society. In terms of the cost performance metric, cost overruns (COs) beyond the initial contract value are more of a norm than a unique occurrence. Therefore, to ensure economic sustainability for SIDS, and value for money on PSIPs, there is a need to investigate and evaluate the risk impacts on COs. The purpose of this research was to identify and evaluate the perceived cost overrun risk factors that are within the primary project stakeholders’ sphere of control, and to reduce the ongoing ambiguities that exist in the prioritization of these risks. This was achieved by extracting critical risk factors from selected comparative studies in developing countries to formulate a closed-ended questionnaire to be administered to construction professionals in Trinidad and Tobago. Thereafter, the process of fuzzy synthetic evaluation (FSE) was used to develop a risk model based on three tiers of risks: 11 critical risk factors, 3 critical risk groupings (CRGs) and an overall risk level (ORL). The results showed that the two highest-ranked critical risks were project funding problems and variations by client. The leading critical risk grouping was client-related risk (5.370), followed by professionalrelated risk (4.815) and physical risk (4.870). The ORL was 5.068. Based on the FSE’s linguistic scaling, the CRGs and the ORL are perceived to be high risks in PSIPs. This research adds to the CO body of knowledge in primarily three ways. Firstly, the study extends the comparative assessment previously undertaken in scholarship into the context of SIDS to build on the generalizability of this context-specific phenomenon. Secondly, the FSE evaluation undertaken provides a practical tool to be promoted for use in SIDS’ construction industry among practitioners to focus and prioritize the critical risks in the planning phases and improve on contemporary risk practices in the execution phases of projects. Finally, this quantitative model approach is recommended to supplement the traditional qualitative risk management practices adopted in SIDS, thus contributing towards the overall improved economic sustainability and viability of PSIPs.Publication Open Access Pavement Roughness Prediction Using Explainable and Supervised Machine Learning Technique for Long-Term Performance(MDPI, 2023-06-15) Sandamal, K; Shashiprabha, S; Muttil, N; Rathnayake, UMaintaining and rehabilitating pavement in a timely manner is essential for preserving or improving its condition, with roughness being a critical factor. Accurate prediction of road roughness is a vital component of sustainable transportation because it helps transportation planners to develop cost-effective and sustainable pavement maintenance and rehabilitation strategies. Traditional statistical methods can be less effective for this purpose due to their inherent assumptions, rendering them inaccurate. Therefore, this study employed explainable and supervised machine learning algorithms to predict the International Roughness Index (IRI) of asphalt concrete pavement in Sri Lankan arterial roads from 2013 to 2018. Two predictor variables, pavement age and cumulative traffic volume, were used in this study. Five machine learning models, namely Random Forest (RF), Decision Tree (DT), XGBoost (XGB), Support Vector Machine (SVM), and K-Nearest Neighbor (KNN), were utilized and compared with the statistical model. The study findings revealed that the machine learning algorithms’ predictions were superior to those of the regression model, with a coefficient of determination (R2) of more than 0.75, except for SVM. Moreover, RF provided the best prediction among the five machine learning algorithms due to its extrapolation and global optimization capabilities. Further, SHapley Additive exPlanations (SHAP) analysis showed that both explanatory variables had positive impacts on IRI progression, with pavement age having the most significant effect. Providing accurate explanations for the decision-making processes in black box models using SHAP analysis increases the trust of road users and domain experts in the predictions generated by machine learning models. Furthermore, this study demonstrates that the use of explainable AI-based methods was more effective than traditional regression analysis in IRI prediction. Overall, using this approach, road authorities can plan for timely maintenance to avoid costly and extensive rehabilitation. Therefore, sustainable transportation can be promoted by extending pavement life and reducing frequent reconstruction.Publication Open Access Wetland Water-Level Prediction in the Context of Machine-Learning Techniques: Where Do We Stand?(MDPI, 2023-05) Jayathilake, T; Gunathilake, M. B; Wimalasiri, E.M; Rathnayake, UWetlands are simply areas that are fully or partially saturated with water. Not much attention has been given to wetlands in the past, due to the unawareness of their value to the general public. However, wetlands have numerous hydrological, ecological, and social values. They play an important role in interactions among soil, water, plants, and animals. The rich biodiversity in the vicinity of wetlands makes them invaluable. Therefore, the conservation of wetlands is highly important in today’s world. Many anthropogenic activities damage wetlands. Climate change has adversely impacted wetlands and their biodiversity. The shrinking of wetland areas and reducing wetland water levels can therefore be frequently seen. However, the opposite can be seen during stormy seasons. Since wetlands have permissible water levels, the prediction of wetland water levels is important. Flooding and many other severe environmental damage can happen when these water levels are exceeded. Therefore, the prediction of wetland water level is an important task to identify potential environmental damage. However, the monitoring of water levels in wetlands all over the world has been limited due to many difficulties. A Scopus-based search and a bibliometric analysis showcased the limited research work that has been carried out in the prediction of wetland water level using machine-learning techniques. Therefore, there is a clear need to assess what is available in the literature and then present it in a comprehensive review. Therefore, this review paper focuses on the state of the art of water-level prediction techniques of wetlands using machine-learning techniques. Nonlinear climatic parameters such as precipitation, evaporation, and inflows are some of the main factors deciding water levels; therefore, identifying the relationships between these parameters is complex. Therefore, machine-learning techniques are widely used to present nonlinear relationships and to predict water levels. The state-of-the-art literature summarizes that artificial neural networks (ANNs) are some of the most effective tools in wetland water-level prediction. This review can be effectively used in any future research work on wetland water-level prediction. © 2023 by the authors.Publication Open Access Sensitivity Analysis of Parameters Affecting Wetland Water Levels: A Study of Flood Detention Basin, Colombo, Sri Lanka(MDPI, 2023-04-02) Herath, M; Jayathilaka, T; Azamathulla, H.M; Mandala, V; Rathnayake, N; Rathnayake, UWetlands play a vital role in ecosystems. They help in flood accumulation, water purification, groundwater recharge, shoreline stabilization, provision of habitats for flora and fauna, and facilitation of recreation activities. Although wetlands are hot spots of biodiversity, they are one of the most endangered ecosystems on the Earth. This is not only due to anthropogenic activities but also due to changing climate. Many studies can be found in the literature to understand the water levels of wetlands with respect to the climate; however, there is a lack of identification of the major meteorological parameters affecting the water levels, which are much localized. Therefore, this study, for the first time in Sri Lanka, was carried out to understand the most important parameters affecting the water depth of the Colombo flood detention basin. The temporal behavior of water level fluctuations was tested among various combinations of hydro-meteorological parameters with the help of Artificial Neural Networks (ANN). As expected, rainfall was found to be the most impacting parameter; however, apart from that, some interesting combinations of meteorological parameters were found as the second layer of impacting parameters. The rainfall–nighttime relative humidity, rainfall–evaporation, daytime relative humidity–evaporation, and rainfall–nighttime relative humidity–evaporation combinations were highly impactful toward the water level fluctuations. The findings of this study help to sustainably manage the available wetlands in Colombo, Sri Lanka. In addition, the study emphasizes the importance of high-resolution on-site data availability for higher prediction accuracy.Publication Open Access Deep Machine Learning-Based Water Level Prediction Model for Colombo Flood Detention Area(MDPI, 2023-02-08) Herath, M; Jayathilaka, T; Hoshino, Y; Rathnayake, UMachine learning has already been proven as a powerful state-of-the-art technique for many non-linear applications, including environmental changes and climate predictions. Wetlands are among some of the most challenging and complex ecosystems for water level predictions. Wetland water level prediction is vital, as wetlands have their own permissible water levels. Exceeding these water levels can cause flooding and other severe environmental damage. On the other hand, the biodiversity of the wetlands is threatened by the sudden fluctuation of water levels. Hence, early prediction of water levels benefits in mitigating most of such environmental damage. However, monitoring and predicting the water levels in wetlands worldwide have been limited owing to various constraints. This study presents the first-ever application of deep machine-learning techniques (deep neural networks) to predict the water level in an urban wetland in Sri Lanka located in its capital. Moreover, for the first time in water level prediction, it investigates two types of relationships: the traditional relationship between water levels and environmental factors, including temperature, humidity, wind speed, and evaporation, and the temporal relationship between daily water levels. Two types of low load artificial neural networks (ANNs) were developed and employed to analyze two relationships which are feed forward neural networks (FFNN) and long short-term memory (LSTM) neural networks, to conduct the comparison on an unbiased common ground. The LSTM has outperformed FFNN and confirmed that the temporal relationship is much more robust in predicting wetland water levels than the traditional relationship. Further, the study identified interesting relationships between prediction accuracy, data volume, ANN type, and degree of information extraction embedded in wetland data. The LSTM neural networks (NN) has achieved substantial performance, including R2 of 0.8786, mean squared error (MSE) of 0.0004, and mean absolute error (MAE) of 0.0155 compared to existing studies.Publication Open Access State-of-the-Art Graphene Synthesis Methods and Environmental Concerns(Hindawi Limited, 2023-02-02) Edward, K; Mamun, K; Narayan, S; Assaf, M; Rohindra, D; Rathnayake, UGraphene, a 2D sp2 hybridized carbon sheet consisting of a honeycomb network, is the building block of graphite. Since its discovery in 2004, graphene's exceptional electronic and mechanical properties have peaked interest in various applications. However, the inability to mass produce high-quality graphene affordably currently limits the practical application of the material. Researchers are continuously working on advancing graphene synthesis methods to alleviate these limitations. Therefore, this review looks at the overview of established graphene synthesis methods and characterization techniques, and then highlights an in-depth review of graphene production through flash joule heating. The environmental concerns related to graphene synthesis are also presented in this review paper.Publication Open Access Spatio-Temporal Rainfall Variability and Concentration over Sri Lanka(Hindawi, 2022-09-28) Pawar, U; Karunathilaka, P; Rathnayake, UChanges in precipitation patterns significantly affect flood and drought hazard management and water resources at local to regional scales. Therefore, the main motivation behind this paper is to examine the spatial and temporal rainfall variability over Sri Lanka by Standardized Rainfall Anomaly Index (SRAI) and Precipitation Concentration Index (PCI) from 1990 to 2019. The Mann–Kendall (MK) trend test and Sen’s slope (SS) were utilized to assess the trend in the precipitation concentration based on PCI. The Inverse Distance Weighting (IDW) interpolation method was incorporated to measure spatial distribution. Precipitation variability analysis showed that seasonal variations are more than those of annual variations. In addition, wet, normal, and dry years were identified over Sri Lanka using SRAI. The maximum SRAI (2.27) was observed for the year 2014 for the last 30 years (1990–2019), which shows the extremely wet year of Sri Lanka. The annual and seasonal PCI analysis showed moderate to irregular rainfall distribution except for the Jaffna and Ratnapura areas (annual scale-positive changes in Katugastota for 21.39% and Wellawaya for 17.6%; seasonal scale-Vavuniya for 33.64%, Trincomalee for 31.26%, and Batticaloa for 18.79% in SWMS). The MK test, SS-test, and percent change analyses reveal that rainfall distribution and concentration change do not show a significant positive or negative change in rainfall pattern in Sri Lanka, despite a few areas which experienced significant positive changes. Therefore, this study suggests that the rainfall in Sri Lanka follows the normal trend of precipitation with variations observed both annually and seasonally.