Showing papers in "Journal of Hydroinformatics in 2013"

Improved annual rainfall-runoff forecasting using PSO-SVM model based on EEMD

Abstract: Rainfall-runoff simulation and prediction in watersheds is one of the most important tasks in water resources management. In this research, an adaptive data analysis methodology, ensemble empirical mode decomposition (EEMD), is presented for decomposing annual rainfall series in a rainfall-runoff model based on a support vector machine (SVM). In addition, the particle swarm optimization (PSO) is used to determine free parameters of SVM. The study data from a large size catchment of the Yellow River in China are used to illustrate the performance of the proposed model. In order to measure the forecasting capability of the model, an ordinary least-squares (OLS) regression and a typical three-layer feed-forward artificial neural network (ANN) are employed as the benchmark model. The performance of the models was tested using the root mean squared error (RMSE), the average absolute relative error (AARE), the coefficient of correlation ( R ) and Nash–Sutcliffe efficiency (NSE). The PSO–SVM–EEMD model improved ANN model forecasting (65.99%) and OLS regression (64.40%), and reduced RMSE (67.7%) and AARE (65.38%) values. Improvements of the forecasting results regarding the R and NSE are 8.43%, 18.89% and 182.7%, 164.2%, respectively. Consequently, the presented methodology in this research can enhance significantly rainfall-runoff forecasting at the studied station.

Real-time urban flood forecasting and modelling – a state of the art

Abstract: All urban drainage networks are designed to manage a maximum rainfall. This situation implies an accepted flood risk for any greater rainfall event. This risk is often underestimated as factors such as city growth and climate change are ignored. But even major structural changes cannot guarantee that urban drainage networks would cope with all future rain events. Thus, being able to forecast urban flooding in real time is one of the main issues of integrated flood risk management. Runoff and hydraulic models can be essential elements of flood forecast systems, as an active part of the system or as studying tools. This paper gives an overview of current available options for pluvial flood modelling in urban areas, from basic estimations with a one-dimensional urban drainage model to detailed flood process representation with one dimensional–two dimensional hydrodynamic coupled models. Each type of modelling solution is described with pros and cons regarding urban flood analysis. The paper then elaborates on real-time flood forecast systems and the influence of their main components. A classification of real-time urban flood systems is given based on the use of urban models, i.e. empirical scenarios, pre-simulated scenarios and real-time simulations. A review of existing operational systems is done using this classification.

Formulation of a fast 2D urban pluvial flood model using a cellular automata approach

Abstract: With the increase in frequency and severity of flash flood events in major cities around the world, the infrastructure and people living in those urban areas are exposed continuously to high risk levels of pluvial flooding. The situation is likely to be exacerbated by the potential impact of future climate change. A fast flood model could be very useful for flood risk analysis. One-dimensional (1D) models provide limited information about the flow dynamics whereas two-dimensional (2D) models require substantial computational time and cost, a factor that limits their use. This paper presents an alternative approach using cellular automata (CA) for 2D modelling. The model uses regular grid cells as a discrete space for the CA setup and applies generic rules to local neighbourhood cells to simulate the spatio-temporal evolution of pluvial flooding. The proposed CA model is applied to a hypothetical terrain and a real urban area. The synchronous state updating rule and inherent nature of the proposed model contributes to a great reduction in computational time. The results are compared with a hydraulic model and good agreement is found between the two models.

Developing reservoir operational decision rule by genetic programming

Abstract: The reservoir operational decision rule is an equation that can balance reservoir system parameters in each period by considering previous experiences of the system. That equation includes variables such as inflow, volume storage and released water from the reservoir that are commonly related to each other by some constant coefficients in predefined linear and nonlinear patterns. Although optimization tools have been extensively applied to develop an optimal operational decision rule, only optimal constant coefficients have been derived and the operational patterns are assumed to be fixed in that operational rule curve. Genetic programming (GP) is an evolutionary algorithm (EA), based on genetic algorithm (GA), which is capable of calculating an operational rule curve by considering optimal operational undefined patterns. In this paper, GP is used to extract optimal operational decision rules in two case studies by meeting downstream water demands and hydropower energy generation. The extracted rules are compared with common linear and nonlinear decision rules, LDR and NLDR, determined by a software package for interactive general optimization (LINGO) and GA. The GP rule improves the objective functions in the training and testing data sets by 2.48 and 8.53%, respectively, compared to the best rule by LINGO and GA in supplying downstream demand. Similarly, the hydropower energy generation improves by 48.03 and 44.21% in the training and testing data sets, respectively. Results show that the obtained objective function value is enhanced significantly for both the training and testing data using GP. They also indicate that the proposed rule, based on GP, is effective in determining optimal rule curves for reservoirs.

Supervised committee machine with artificial intelligence for prediction of fluoride concentration

Abstract: The study introduces a supervised committee machine with artificial intelligence (SCMAI) method to predict fluoride in ground water of Maku, Iran. Ground water is the main source of drinking water for the area. Management of fluoride anomaly needs better prediction of fluoride concentration. However, the complex hydrogeological characteristics cause difficulties to accurately predict fluoride concentration in basaltic formation, non-basaltic formation, and mixing zone. SCMAI predicts fluoride by a nonlinear combination of individual AI models through an artificial intelligent system. Factor analysis is used to identify effective fluoride-correlated hydrochemical parameters as input to AI models. Four AI models, Sugeno fuzzy logic, Mamdani fuzzy logic, artificial neural network (ANN), and neuro-fuzzy are employed to predict fluoride concentration. The results show that all of these models have similar fitting to the fluoride data in the Maku area, and do not predict well for samples in the mixing zone. The SCMAI employs an ANN model to re-predict the fluoride concentration based on the four AI model predictions. The result shows improvement to the CMAI method, a committee machine with the linear combination of AI model predictions. The results also show significant fitting improvement to individual AI models, especially for fluoride prediction in the mixing zone.

Conjunction of SOM-based feature extraction method and hybrid wavelet-ANN approach for rainfall–runoff modeling

Abstract: In rainfall–runoff modeling, the wavelet-ANN model, which includes a wavelet transform to capture multi-scale features of the process, as well as an artificial neural network (ANN) to predict the runoff discharge, is a beneficial approach. One of the essential steps in any ANN-based development process is determination of dominant input variables. This paper presents a two-stage procedure to model the rainfall–runoff process of the Delaney Creek and Payne Creek Basins, Florida, USA. The two-stage procedure includes data pre-processing and model building stages. In the data pre-processing stage, a wavelet transform is used to decompose the rainfall and runoff time series into several sub-series at different scales. Subsequently, independent sub-series are chosen via a self-organizing map (SOM). In the model building stage, selected sub-series are imposed as input data to a feed-forward neural network (FFNN) to forecast runoff discharge. To make a better interpretation of the model efficiency, the proposed model is compared with the Auto Regressive Integrated Moving Average with eXogenous input (ARIMAX) and with the ad hoc FFNN methods, without any data pre-processing. The results proved that the proposed model leads to better outcome especially in term of determination coefficient for detecting peak points (DCpeak).

Exploring the potential of SRTM topographic data for flood inundation modelling under uncertainty

Abstract: The desirable data for model building and calibration to support the decision-making process in flood risk management are often not sufficient or unavailable. A potential opportunity is now offered by global remote sensing data, which can be freely (or at low cost) obtained from the internet, for example, Shuttle Radar Topography Mission (SRTM) topography. There is a general sense that inundation modelling performance will be degraded by using SRTM topography data. However, the actual effectiveness and usefulness of SRTM topography is still largely unexplored. To overcome this lack of knowledge, we have explored the value of SRTM topography to support flood inundation modelling under uncertainty. The study was performed on a 98 km reach of the River Po in northern Italy. The comparison between a hydraulic model based on high-quality topography and one based on SRTM topography was carried out by explicitly considering other sources of uncertainty (besides topography inaccuracy) that unavoidably affect hydraulic modelling, such as parameter and inflow uncertainties. The results of this study showed that the differences between the high-resolution topography-based model and the SRTM-based model are significant, but within the accuracy that is typically associated with large-scale flood studies.

Information theory applied to evaluate the discharge monitoring network of the Magdalena River

Abstract: The acquisition of good hydrologic information is an important issue in water management since it is the basis of decisions concerning the allocation of water resources to different users. However, sufficient data are often not available to describe the behaviour of such systems, especially in developing countries, where monitoring networks are inappropriately designed, poorly operated or are inadequate. Therefore, it is of interest to design and evaluate efficient monitoring networks. This paper presents two methodologies to design discharge monitoring networks in rivers using the concepts of Information Theory. The first methodology considers the optimization of Information Theory quantities and the second considers a new method that is based on ranking Information Theory quantities with different possible monitor combinations. The methodologies are tested for the Magdalena River in Colombia, in which the existing monitoring network is also assessed. In addition, the use of monitors at tributaries is explored. It is demonstrated that the ranking method is a promising way of finding the extremes of Pareto fronts generated during multi-objective optimization processes and that better (more informative and less redundant) monitoring network configurations can be found for the Magdalena River.

Geostatistical techniques for approximate location of pipe burst events in water distribution systems

Abstract: This paper focusses on the customisation and further enhancement of the recently developed data-driven methodology for the automated near real-time detection of pipe bursts and other (e.g. sensor faults) events at the district metered area (DMA) level. Assuming the availability of pressure/flow data from an increased number of sensors deployed in a DMA, the aim is to: (i) overcome the limitations of the probabilistic inference engine when dealing with the increased data availability; and (ii) exploit the event information resulting from the analysis of the larger number of DMA signals for determining the approximate location of the pipe burst events within the DMA. This is achieved by making use of a multivariate Gaussian mixtures-based graphical model and geostatistical techniques. The novel detection and location methodology is demonstrated and tested on a series of simulated pipe burst events that were performed by opening hydrants in a real-life DMA in the UK. The results obtained illustrate that the new methodology can successfully determine the approximate location of pipe bursts within a DMA (in addition to detecting them in a fast and reliable manner). The performance comparison of several geostatistical techniques shows that the Ordinary Cokriging technique outperforms all other techniques tested.

Cloud and cluster computing in uncertainty analysis of integrated flood models

Abstract: There is an increased awareness of the importance of flood management aimed at preventing human and material losses. A wide variety of numerical modelling tools have been developed in order to make decision-making more efficient, and to better target management actions. Hydroinformatics assumes the holistic integrated approach to managing the information propagating through models, and analysis of uncertainty propagation through models is an important part of such studies. Many popular approaches to uncertainty analysis typically involve various strategies of Monte Carlo sampling of uncertain variables and/or parameters and running a model a large number of times, so that in the case of complex river systems this procedure becomes very time-consuming. In this study the popular modelling systems HEC-HMS, HEC-RAS and Sobek1D2D were applied to modelling the hydraulics of the Timis–Bega basin in Romania. We considered the problem of studying how the flood inundation is influenced by uncertainties in water levels of the reservoirs in the catchment, and uncertainties in the digital elevation model (DEM) used in the 2D hydraulic model. For this we used cloud computing (Amazon Elastic Compute Cloud platform) and cluster computing on the basis of a number of office desktop computers, and were able to show their efficiency, leading to a considerable reduction of the required computer time for uncertainty analysis of complex models. The conducted experiments allowed us to associate probabilities to various areas prone to flooding. This study allows us to draw a conclusion that cloud and cluster computing offer an effective and efficient technology that makes uncertainty-aware modelling a practical possibility even when using complex models.

An agent-based modeling framework for sociotechnical simulation of water distribution contamination events

Abstract: In the event that a contaminant is introduced to a water distribution network, a large population of consumers may risk exposure. Selecting mitigation actions to protect public health may be difficult, as contamination is a poorly predictable dynamic event. Consumers who become aware of an event may select protective actions to change their water demands from typical demand patterns, and new hydraulic conditions can arise that differ from conditions that would be predicted when demands are considered as exogenous inputs. Consequently, the movement of the contaminant plume in the pipe network may shift from its expected trajectory. A sociotechnical model is developed here to integrate agent-based models of consumers with an engineering water distribution system model and capture the dynamics between consumer behaviors and the water distribution system for predicting contaminant transport and public exposure. Consumers are simulated as agents with behaviors, including movement, water consumption, exposure, reduction in demands, and communication with other agents. As consumers decrease their water use, the location of the contaminant plume is updated and the amount of contaminant consumed by each agent is calculated. The framework is tested through simulating realistic contamination scenarios for a virtual city and water distribution system.

Extension of the constrained ant colony optimization algorithms for the optimal operation of multi-reservoir systems

Abstract: This paper extends the application of Constrained Ant Colony Optimization Algorithms (CACOAs) to optimal operation of multi-reservoir systems. Three different formulations of the constrained Ant Colony Optimization (ACO) are outlined here using Max-Min Ant System for the solution of multi-reservoir operation problems. In the first two versions, called Partially Constrained ACO algorithms, the constraints of the multi-reservoir operation problems are satisfied partially. In the third formulation, all the constraints of the underlying problem are implicitly satisfied by the provision of tabu lists to the ants which contain only feasible options. The ants are, therefore, forced to construct feasible solutions and hence the method is referred to as a Fully Constrained ACO algorithm. The proposed constrained ACO algorithms are formulated for both possible cases of taking storage/release volumes as the decision variables of the problem. The proposed methods are used to optimally solve the well-known problems of four- and ten-reservoir operations and the results are presented and compared with those of the conventional unconstrained ACO algorithm and existing methods in the literature. The results indicate the superiority of the proposed methods over conventional ACOs and existing methods to optimally solve large scale multi-reservoir operation problems.

Distributed tree-based model predictive control on a drainage water system

Abstract: Open water systems are one of the most externally influenced systems due to their size and continuous exposure to uncertain meteorological forces. The control of systems under uncertainty is, in general, a challenging problem. In this paper, we use a stochastic programming approach to control a drainage system in which the weather forecast is modeled as a disturbance tree. Each branch of the tree corresponds to a possible disturbance realization and has a certain probability associated to it. A model predictive controller is used to optimize the expected value of the system variables taking into account the disturbance tree. This technique, tree-based model predictive control (TBMPC), is solved in a distributed fashion. In particular, we apply dual decomposition to get an optimization problem that can be solved by different agents in parallel. In addition, different possibilities are considered in order to reduce the communicational burden of the distributed algorithm without reducing the performance of the controller significantly. Finally, the performance of this technique is compared with others such as minmax or multiple MPC.

Numerical study of 3D flow at right-angled confluences with and without upstream planform curvature

Abstract: The purpose of this paper is to study the individual and combined effect of upstream planform curvature and difference in bed elevations at the tributary entrance to the confluence on the flow in the confluence hydrodynamics zone. To do this, flow at right-angled confluences with three planforms and four values of bed elevation discordance ratio (Δ z T/ h d) is simulated using a three-dimensional (3D) numerical model. Three confluence planforms include confluences with the (1) straight tributary canal (SC), (2) right bend (RB) and (3) left bend (LB) in the tributary. Four Δ z T/ h d values in the range [0.0, 0.5] include both concordant and discordant beds' confluences. Overall, nine cases with the straight main canal are considered. Special attention is paid to the flow deflection and flow separation zones since the former affects transfer of momentum from the tributary to the main canal and the latter affects transport capacity of the post-confluence channel. Comparison of the results reveals that the influence of RB in the tributary is practically negligible in comparison to the straight canal case. With the increasing difference in bed elevations between the tributary and main canals (Δ z T), the presence of LB strengthens 3D flow and the structure of the recirculation zone is destroyed.

Artificial neural network (ANN) modeling of dynamic effects on two-phase flow in homogenous porous media

Abstract: The dynamic effect in two-phase flow in porous media indicated by a dynamic coefficient τ depends on a number of factors (e.g. medium and fluid properties). Varying these parameters parametrically in mathematical models to compute τ incurs significant time and computational costs. To circumvent this issue, we present an artificial neural network (ANN)-based technique for predicting τ over a range of physical parameters of porous media and fluid that affect the flow. The data employed for training the ANN algorithm have been acquired from previous modeling studies. It is observed that ANN modeling can appropriately characterize the relationship between the changes in the media and fluid properties, thereby ensuring a reliable prediction of the dynamic coefficient as a function of water saturation. Our results indicate that a double-hidden-layer ANN network performs better in comparison to the single-hidden-layer ANN models for the majority of the performance tests carried out. While single-hidden-layer ANN models can reliably predict complex dynamic coefficients (e.g. water saturation relationships) at high water saturation content, the double-hidden-layer neural network model outperforms at low water saturation content. In all the cases, the single- and double-hidden-layer ANN models are better predictors in comparison to the regression models attempted in this work.

Comparison of three global optimization algorithms for calibration of the Xinanjiang model parameters.

Abstract: The Xinanjiang model, a conceptual rainfall-runoff (CRR) model with distributed parameters, has been successfully and widely applied to flood forecasting of large basins in humid and semi-humid regions of China. With an increasing demand for timely and accurate forecasts in hydrology, how to obtain more appropriate parameters for CRR models has long been an important topic. These models have a large number of parameters which cannot be directly obtained from measurable quantities of catchments characteristics. In this study, three different optimization methods are used to calibrate the Xinanjiang streamflow model: genetic algorithm (GA), shuffled complex evolution of the University of Arizona (SCE-UA) and the recently developed shuffled complex evolution Metropolis algorithm of the University of Arizona (SCEM-UA), using streamflow data of the Shuangpai Reservoir in China. Two different time steps of 1 and 3 hr are used in the analysis. The results indicate that the SCEM-UA algorithm can infer the most probable parameter set and furnish useful information about the nature of the response surface in the vicinity of the optimum. Moreover, there is larger uncertainty for 1 hr forecasting than for 3 hr forecasting. This is significant in assessing risks in likely applications of Xinanjiang models.

Numerical assessment of bed-load discharge formulations for transient flow in 1D and 2D situations

Abstract: Two-dimensional (2D) transient flow over an erodible bed can be modelled using shallow-water equations and the Exner equation to describe the morphological evolution of the bed. Considering the fact that well-proven capacity formulae are based on one-dimensional (1D) experimental steady flows, the assessment of these empirical relations under unsteady 1D and 2D situations is important. In order to ensure the reliability of the numerical experimentation, the formulation has to be general enough to allow the use of different empirical laws. Moreover, the numerical scheme must handle correctly the coupling between the 2D shallow-water equations and the Exner equation under any condition. In this work, a finite-volume numerical scheme that includes these two main features will be exploited here in 1D and 2D laboratory test cases. The relative performances of Meyer-Peter and Muller, Ashida and Michiue, Engelund and Fredsoe, Fernandez Luque and Van Beek, Parker, Smart, Nielsen, Wong and Camenen and Larson formulations are analysed in terms of the root mean square error. A new discretization of the Smart formula is provided, leading to promising predictions of the erosion/deposition rates. The results arising from this work are useful to justify the use of an empirical sediment bed-load discharge formula among the ones studied, regardless of the hydrodynamic situation.

Calibration of a 2-D morphodynamic model using water–sediment flux maps derived from an ADCP recording

Abstract: This study compared numerical simulations of the hydro-morphodynamics of a stretch of the Po River (Italy) to detailed measurements of the river’s morphology and water–sediment fluxes. A survey of the 8-km-long, 250-m-wide river section, previously described in former publications, was performed using two vessels. The first was equipped with a multibeam echo sounder for bathymetry and the investigation of bed forms, while the second was equipped with two acoustic Doppler current profilers (ADCPs) to map bed roughness, flow velocity, and suspended sediment fields. This paper focuses on the calibration of the shallow water numerical model MIKE21C implemented by the Danish Hydraulic Institute. Sensitivity analyses were performed to analyze the effects of bed roughness and sediment transport direction on the simulated flow field and morphology. The results fixed the predominant flow and morphology parameters in the model, and show how the velocity and sediment field maps that were derived from the ADCP recording could be used for calibration and validation of the existing software.

Improving reliability of river flow forecasting using neural networks, wavelets and self-organising maps

Abstract: Neural network (NN) models have gained much attention for river flow forecasting because of their ability to map complex non-linearities However, the selection of appropriate length of training datasets is crucial and the uncertainty in predictions of the trained NNs with new datasets is a crucial problem In this study, self-organising maps (SOM) are used to classify the datasets homogeneously and the performance of four types of NN models developed for daily discharge predictions – namely traditional NN, wavelet-based NN (WNN), bootstrap-based NN (BNN) and wavelet-bootstrap-based NN (WBNN) – is analysed for their applicability cluster-wise SOM classified the training datasets into three clusters (ie cluster I, II and III) and the trained SOM is then used to assign testing datasets into these three clusters Simulation studies show that the WBNN model performs better for the entire testing dataset as well as for values in clusters I and III; for cluster II the performance of BNN model is better compared with others for a 1-day lead time forecasting Overall, it is found that the proposed methodology can enhance the accuracy and reliability of river flow forecasting

Estimation of current-induced pile groups scour using a rule-based method

Abstract: Scour phenomenon around piles could endanger the stability of the structures placed on them. Therefore, an accurate estimation of the scour depth around piles is very important for engineers. Due to the complexity of the interaction between the current, seabed and pile group; prediction of the scour depth is a difficult task and the available empirical formulas have limited accuracy. Recently, soft computing methods such as artificial neural networks (ANN) and adaptive neuro-fuzzy inference system (ANFIS) have been used for the prediction of the scour depth. However, these methods do not give enough insight into the generated models and are not as easy to use as the empirical formulas. In this study, new formulas are given that are compact, accurate and physically sound. In comparison with the other soft computing methods, this approach is more transparent and robust. Comparison between the developed formulas and previous empirical formulas showed the superiority of the developed ones in terms of accuracy. In addition, the given formulas can be easily used by engineers to estimate the scour depth around pile groups. Moreover, in this study, design factors are given for different levels of acceptable risks, which can be useful for design purposes.

An evaluation of ANN methods for estimating the lengths of hydraulic jumps in U-shaped channel

Abstract: Modelling of hydraulic characteristics of jump using theoretical and empirical models has always been a difficult task. The length of jump may be defined as the distance measured from the toe of the jump to the location of the surface rise. Due to high turbulence this length cannot be determined easily by theory. However, it has been investigated experimentally so as to design the stilling basins with hydraulic jumps. In this work, the control of a hydraulic jump by broad-crested sills in a U-shaped channel is recalled theoretically and experimentally examined. The study begins with a multiple regression (MR) analysis. Then, and in order to model the relative lengths of hydraulic jumps, we have implemented and evaluated two different artificial neural networks (ANN): multilayer perceptron neural network (MLPNN) and generalized regression neural network (GRNN). The results demonstrate the predictive strength of GRNN and its potential to predict hydraulic problems with an adaptive spread value. However, the MLPNN model remains best classified by these indexes of performance.

Wavelets-based non-linear model for real-time daily flow forecasting in Krishna River

Abstract: In this study, a multi-scale non-linear model based on coupling a discrete wavelet transform (DWT) and the second-order Volterra model, i.e. the wavelet Volterra coupled (WVC) model, is applied for daily inflow forecasting at Krishna Agraharam, Krishna River, India. The relative performance of the WVC model was compared with regular artificial neural networks (ANN), wavelet-artificial neural networks (WA-ANN) models and other baseline models such as auto-regressive moving average with exogenous variables (ARMAX) for lead times of 1–5 days. The models were applied for the forecasting of daily streamflow at Krishna Agraharam Station at Krishna River. The WVC performed very well, especially when compared with the WA-ANN model for lead times of 4 and 5 days. The results indicate that the WVC model is a promising alternative to the other traditional models for short-term flow forecasting.

Analysis of agricultural pollution by flood flow impact on water quality in a reservoir using a three-dimensional water quality model

Abstract: This study presents the Yuqiao Reservoir Water Quality Model (YRWQM), a three-dimensional hydrodynamic and water quality model of the Yuqiao reservoir, China. The YRWQM was developed under the environmental fluid dynamics code (EFDC) model and was calibrated and verified to hydrodynamic and water quality data, using two sets of observed data from January 1 to December 31, 2006 and from May 1 to October 31, 2007, respectively. The primary hydrodynamic and transport driving forces are inflows/outflows and surface wind stresses. Considering effects of water transfer and wind on the advection-dispersion processes, the model results showed better agreements with observed data in the reservoir. The YRWQM predicted the variations of water quality resulting from agricultural pollution which flowed into the reservoir with floods lasting for 12 days in 2009. The results indicated that the concentrations of chemical oxygen demand and total nitrogen were increased 225 and 314%, respectively. Considering the interactions between chlorophyll- a and nitrogen in the model, the results indicated the reservoir was not a nitrogen-limited environment. We suggest the management should focus on agricultural pollution strategies for the reservoir during the flood period. The YRWQM could be a useful tool for water sources management in the reservoir.

Modelling floods in urban areas and representation of buildings with a method based on adjusted conveyance and storage characteristics

Abstract: The present paper reviews several approaches that can be used in capturing urban features in coarse resolution two-dimensional (2D) models and it demonstrates the effectiveness of a new approach against the straightforward 2D modelling approach on a hypothetical and a real-life case study work. The case study work addresses the use of coarse grid resolutions in 2D non-inertia models. The 2D non-inertia model used solves continuity and momentum equations over the cells of the coarse model while taking the minimum elevation as a surface level. The volume stored in every cell is calculated as a volume-depth relationship. In order to replicate restriction in conveyances in x – y directions of fine resolution models due to building blocks, the friction values of the coarse-resolution model are adjusted to match the results of the high-resolution model. The work presented in this paper shows the possibility of applying a 2D non-inertia model more effectively in urban flood modelling applications whilst still making use of the high resolution of topographic data that can nowadays be easily acquired.

Flood wave propagation in steep mountain rivers

Abstract: Most of the recent developments concerning efficient numerical schemes to solve the shallow-water equations in view of real world flood modelling purposes concern the two-dimensional form of the equations or the one-dimensional form written for rectangular, unit-width channels. Extension of these efficient schemes to the one-dimensional cross-sectional averaged shallow-water equations is not straightforward, especially when complex natural topographies are considered. This paper presents different formulations of numerical schemes based on the HLL (Harten-Lax-van Leer) solver, and the adaptation of the topographical source term treatment when cross-sections of arbitrary shape are considered. Coupled and uncoupled formulations of the equations are considered, in combination with centred and lateralised source term treatment. These schemes are compared to a numerical solver of Lax Friedrichs type based on a staggered grid. The proposed schemes are first tested against two theoretical benchmark tests and then applied to the Brembo River, an Italian alpine river, firstly simulating a steady-state condition and secondly reproducing the 2002 flood wave propagation. © IWA Publishing 2013.

Sensitivity analysis of bridge pier scour depth predictive formulae

Abstract: Sensitivity analysis is an approach to recognising the behaviour of models and relative importance of causative factors. In this paper, behaviours of six pier scour depth empirical formulae are evaluated on the basis of an analytical method. The sensitivity of predicted scour depth is analysed with respect to the following independent parameters: approach flow depth, riverbed slope and median sediment size. Also their combined influence is studied examining the relative importance of each parameter with respect to the total variation of the maximum scour depth. Results show that: (1) sensitivity significantly depends on flow intensity for most of the selected formulae, whereas for the others it is a constant value or depends on other influencing parameters; (2) different formulae demonstrate various level of sensitivity to the input variables, so that, for a certain error in the input variables, the error in the results may vary consistently; (3) some formulae are very sensitive to the input parameters under some conditions, hence an error in an input variable may be amplified in the output results; and (4) most of the formulae are more sensitive to the variations of the influencing parameters in clear-water than in live-bed conditions.

Tree-based fitted Q-iteration for multi-objective Markov decision processes in water resource management

Abstract: Multi-objective Markov decision processes (MOMDPs) provide an effective modeling framework for decision-making problems involving water systems. The traditional approach is to define many single-objective problems (resulting from different combinations of the objectives), each solvable by standard optimization. This paper presents an approach based on reinforcement learning (RL) that can learn the operating policies for all combinations of objectives in a single training process. The key idea is to enlarge the approximation of the action-value function, which is performed by single-objective RL over the state-action space, to the space of the objectives9 weights. The batch-mode nature of the algorithm allows for enriching the training dataset without further interaction with the controlled system. The approach is demonstrated on a numerical test case study and evaluated on a real-world application, the Hoa Binh reservoir, Vietnam. Experimental results on the test case show that the proposed approach (multi-objective fitted Q-iteration; MOFQI) becomes computationally preferable over the repeated application of its single-objective version (fitted Q-iteration; FQI) when evaluating more than five weight combinations. In the Hoa Binh case study, the operating policies computed with MOFQI and FQI have comparable efficiency, while MOFQI provides a continuous approximation of the Pareto frontier with no additional computing costs.

A general multi-objective hyper-heuristic for water distribution network design with discolouration risk

Abstract: The optimisation of water distribution networks (WDNs) by evolutionary algorithms has gained much coverage in the literature since it was first proposed in the early 1990s. Despite being well studied, the problem and objectives continue to evolve as demands on water companies change. Motivated by the increased focus on reducing the risk of discolouration, this study examines a three objective version of the WDN design problem which takes into account cost, head excess and discolouration risk. Using this formulation, this paper presents a method for producing optimised network designs aimed at reducing discolouration risk in the network design phase and thus reducing the associated long-term maintenance and operational burdens of the system. This paper discusses the use of a discolouration risk model and, using this model, the optimisation of network design, specifically pipe diameters, to produce a range of high quality self-cleaning networks. The network designs are optimised using the Markov-chain hyper-heuristic (MCHH), a new multi-objective online selective hyper-heuristic. The MCHH is incorporated in to the known NSGA-II and SPEA2 and supplied with a range of heuristics tailored for use on the WDN design problem. The results demonstrate an improvement in performance obtained over the original algorithms.

Development of a real-time regional-inundation forecasting model for the inundation warning system

Abstract: Accurate forecasts of the inundation depth are necessary for inundation warning and mitigation. In this paper, a real-time regional forecasting model is proposed to yield 1- to 3-h lead time inundation maps. First, the K -means based cluster analysis is developed to group the inundation depths and to indentify the control points. Second, the support vector machine is used as the computational method to develop the point forecasting module to yield inundation forecasts for each control point. Third, based on the forecasted depths and the geographic information, the spatial expansion module is developed to expand the point forecasts to the spatial forecasts. An actual application to Siluo Township, Taiwan, is conducted to demonstrate the advantage of the proposed model. The results indicate that the proposed model can provide accurate inundation maps for 1- to 3-h lead times. The accurate long lead time forecasts can extend the lead time to allow sufficient time to take emergency measures. Furthermore, the proposed model is an efficient process that can be trained rapidly with real-time data and is more suitable to be integrated with the decision support system. In conclusion, the proposed modeling technique is expected to be useful to support the inundation warning systems.

Experimental investigation of leak hydraulics

Abstract: In recent decades the hydraulics of leaks, i.e. the definition of the relationships linking the hydraulic quantities in pipes with leaks, has received increasing attention. On the one hand, the definition of the relationship between the leak outflow and the relevant parameters – e.g. the leak area and shape, the pressure inside the pipe and outside the leak, and the pipe material – is crucial for pressure control and inverse analysis techniques. On the other hand, if the effect of the leakage on the governing equations is not taken into account, i.e. the loss of the flow axial momentum is not considered, significant errors can be introduced in the simulation of water distribution systems. In this paper, the governing equations for a pipe with a leak are derived. The basic equations, obtained within different approaches, are presented in a consistent formulation and then compared with the results of some experimental tests. The leak jet angle and other major features of the results are analysed. The estimated values of the parameters can be used in the water distribution network models when pipes with a diffuse leakage are considered.