Didier Graillot

Application of Artificial Neural Networks and Particle Swarm Optimization for the Management of Groundwater Resources

Ground management problems are typically solved by the simulation-optimization approach where complex numerical models are used to simulate the groundwater flow and/or contamination transport. These numerical models take a lot of time to solve the management problems and hence become computationally expensive. In this study, Artificial Neural Network (ANN) and Particle Swarm Optimization (PSO) models were developed and coupled for the management of groundwater of Dore river basin in France. The Analytic Element Method (AEM) based flow model was developed and used to generate the dataset for the training and testing of the ANN model. This developed ANN-PSO model was applied to minimize the pumping cost of the wells, including cost of the pipe line. The discharge and location of the pumping wells were taken as the decision variable and the ANN-PSO model was applied to find out the optimal location of the wells. The results of the ANN-PSO model are found similar to the results obtained by AEM-PSO model. The results show that the ANN model can reduce the computational burden significantly as it is able to analyze different scenarios, and the ANN-PSO model is capable of identifying the optimal location of wells efficiently.

Combined use of groundwater modeling and potential zone analysis for management of groundwater

A methodology for groundwater evaluation has been developed by the combined use of numerical model and spatial modeling using GIS. The developed methodology has been applied on the sub-basin of the Banganga River, India. Initially, the groundwater potential zones have been delineated by spatial modeling. Different thematic maps of the basin like geology, geomorphology, soil, drainage, slope factor and landuse/landcover have been used to identify the groundwater potential zones. Further, the groundwater flow model for the study area has been developed in the MODFLOW. The groundwater flow vector map has been developed and superimposed on the potential zone map to validate the results of spatial modeling. Finally, the different scenarios have been conceptualized by varying the discharge of the wells and purposing the location for new rainwater harvesting structures. Results reveal that increasing the discharge of the wells in the potential zones put less stress on the aquifer. The suggested locations of rainwater harvesting structures also help to reduce the overall decline of groundwater in the area. The hydrological and spatial modeling presented in this study is highly useful for the evaluation of groundwater resources and for deciding the location of rainwater harvesting structures in semi-arid regions.

Storm-water management through Infiltration trenches

With urbanization, the permeable soil surface area through which recharge by infiltration can occur is reducing. This is resulting in much less ground-water recharge and greatly increased surface run-off. Infiltration devices, which redirect run-off waters from the surface to the sub-surface environments, are commonly adopted to mitigate the negative hydrologic effects associated with urbanization. An infiltration trench alone or in combination with other storm water management practice is a key element in present day sustainable urban drainage systems. A solution for the infiltration rate from an infiltration trench and, consequently, the time required to empty the trench is presented. The solution is in the form of integrals of complicated functions and requires numerical computation. The solution is useful in quantifying infiltration rate and/or artificial recharge of ground-water through infiltration trenches and the drain time of the trench, which is a key parameter in operation of storm water manage...

Fault detection on a sewer network by a combination of a Kalman filter and a binary sequential probability ratio test

Abstract One of the aims of sewer networks is the protection of population against floods and the reduction of pollution rejected to the receiving water during rainy events. To meet these goals, managers have to equip the sewer networks with and to set up real-time control systems. Unfortunately, a component fault (leading to intolerable behaviour of the system) or sensor fault (deteriorating the process view and disturbing the local automatism) makes the sewer network supervision delicate. In order to ensure an adequate flow management during rainy events it is essential to set up procedures capable of detecting and diagnosing these anomalies. This article introduces a real-time fault detection method, applicable to sewer networks, for the follow-up of rainy events. This method consists in comparing the sensor response with a forecast of this response. This forecast is provided by a model and more precisely by a state estimator: a Kalman filter. This Kalman filter provides not only a flow estimate but also an entity called ‘innovation’. In order to detect abnormal operations within the network, this innovation is analysed with the binary sequential probability ratio test of Wald. Moreover, by crossing available information on several nodes of the network, a diagnosis of the detected anomalies is carried out. This method provided encouraging results during the analysis of several rains, on the sewer network of Seine-Saint-Denis County, France.

Coupling groundwater modeling and biological indicators for identifying river/aquifer exchanges

Future climate changes and the resulting modifications in anthropogenic activities will alter the interactions between rivers and groundwater. The quantification of these hydraulic interactions is absolutely necessary for achieving sustainable water use and requires accurate analytical methodologies. This report proposes an interdisciplinary approach to the quantitative and qualitative characterization of hydraulic interactions between rivers and shallow aquifers, wherein it outlines the advantages of coupling groundwater modeling with biological markers. As a first step, we built independent diagnostic maps of hydrological exchanges at the sector scale on the basis of hydrogeological modeling and biological indicators. In a second step, these maps were compared to provide a quantitative and qualitative understanding of exchanges between groundwater and surface water. This comparison significantly improved the calibration of groundwater models through a better assessment of boundary zones. Our approach enabled us to identify the conditions under which it could be possible to use biological indicators instead of a large set of piezometric measures. The integration of such combined tools in a future decision support system will assist governmental authorities in proposing appropriate long-term water policies for the preservation of groundwater resources, such as for supplying potable water and/or mitigating pollution risks.

Brief paper: Knowledge-based real-time fault detection and supervision of urban drainage systems

The paper presents the development of a real-time expert system (RTES) for supervision and fault detection in urban drainage systems (UDS). Several vital characteristics of such an expert tool, in which the project is being developed, are recalled. The main aspects of the modelling of the UDS and the detection of faults are outlined and illustrated for the case of a test urban area, north-east of Paris. Procedural knowledge about the behaviour of the UDS is implemented using formulae and procedures. Object-oriented representation enables structuration and formulation of the UDS structure. Rule bases containing expert, heuristic knowledge are written in a quasi-natural language. Knowledge formulation and realtime implementation are illustrated on a RTES-prototype designed with G2 (Gensym Corp.).

Prediction of ground water levels in the uplands of a tropical coastal riparian wetland using artificial neural networks

Artificial Neural Networks (ANNs) have been found to be a robust tool to model many non-linear hydrological processes. The present study aims at evaluating the performance of ANN in simulating and predicting ground water levels in the uplands of a tropical coastal riparian wetland. The study involves comparison of two network architectures, Feed Forward Neural Network (FFNN) and Recurrent Neural Network (RNN) trained under five algorithms namely Levenberg Marquardt algorithm, Resilient Back propagation algorithm, BFGS Quasi Newton algorithm, Scaled Conjugate Gradient algorithm, and Fletcher Reeves Conjugate Gradient algorithm by simulating the water levels in a well in the study area. The study is analyzed in two cases-one with four inputs to the networks and two with eight inputs to the networks. The two networks-five algorithms in both the cases are compared to determine the best performing combination that could simulate and predict the process satisfactorily. Ad Hoc (Trial and Error) method is followed in optimizing network structure in all cases. On the whole, it is noticed from the results that the Artificial Neural Networks have simulated and predicted the water levels in the well with fair accuracy. This is evident from low values of Normalized Root Mean Square Error and Relative Root Mean Square Error and high values of Nash-Sutcliffe Efficiency Index and Correlation Coefficient (which are taken as the performance measures to calibrate the networks) calculated after the analysis. On comparison of ground water levels predicted with those at the observation well, FFNN trained with Fletcher Reeves Conjugate Gradient algorithm taken four inputs has outperformed all other combinations.

Multi-Criteria Decision Analysis for Identifying a Suitable Location for Groundwater Pumping Wells

The paper presents the methodology for the combined use of GIS-based multi-criteria analysis and simulation-optimisation modelling for management of the groundwater resources of the Dore river basin in France. The study identifies the suitable location and maximum discharge for the new groundwater pumping wells. The multi-criteria analysis (MCA), with the help of GIS-based geospatial analysis, was performed to identify those areas suitable for pumping wells by considering different criteria, such as hydraulic conductivity, land use, river aquifer exchange, depth to water, and geomorphology. Different criteria were selected with the help of regional experts and stakeholders. For the study area, the groundwater flow model was developed. Further, new pumping wells in the suitable zones, those identified by MCA, were considered and a simulation-optimisation technique was used to identify the maximum discharge from those wells. Finally, the results obtained from both the methods were to finalise the potential zone. The developed methodology proves to be a more realistic approach to identifying new locations for pumping wells.

Fate and effect of pollutants in rivers: from analysis to modeling

The urban, agricultural and industrial developments generate many pollutant emissions, in increasing quantities and of a highly variable nature, in aquatic environments (nutrients, pesticide residues, heavy metals, pharmaceutical residues...). This global statement covers notable differences between countries. On the one hand, the sometimes alarming river pollution levels in developing countries (hyper eutrophication, toxic substances...) are mainly due to the lack or low efficiency of waste treatment plants, and on the other hand, the socalled developed industrialized countries have established efficient wastewater treatment infrastructures, with restrictive regulations (e.g., european Water Framework Directive, WFD) and technical and financial resources to carry out recovery plan actions. In addition to chemical releases, other pressures also impact the river’s quality: suspended solid and sediment from land erosion, hypoxia (or anoxia) during the microbial degradation of algal blooms or of organic matter, reduction of flows increasing pollutant concentrations, and increase in water temperature. Lastly, the many interactions between contaminants/pollutants and environmental factors control the expected stresses and their effects on water uses and aquatic biodiversity. Aquatic pollution—whether widespread or occasional and chronic or accidental—is often the result of a complex mixture of substances and environmental factors, making risk assessment particularly difficult and increasing uncertainty about the forecasted effects. With the exception of heavy metals, organic pollutants or toxic compounds are degradable, sooner or later, due to the various natural attenuation processes: photodegradation, biodegradation... These processes are however likely to affect the bioavailability and toxicity of substances and globally reduce the toxic load over time. But for this to happen, the biodegradation potential of the rivermust not be exceeded by excessive pollutant loads, which is often the case given the continuous inputs along the river and the upstream-downstream transfer of pollution: a watercourse is not and should never be considered as an annex of a wastewater treatment plant. The induced environmental effects are also multiple and can be evaluated at different overlapping degrees, making a meaningful assessment of in situ induced effects extremely complicated: from the individual to communities/acute or chronic effects, direct or indirect (domino effect); physiological and/or genetic effects (multigenerational and transmittable)... From a scientific point of view, all these issues are brought together within a thematic/new discipline called stress ecology, making it possible to formalize pressure-impact relationships, whatever the causes of environmental stress are. Numerous methods of bioindication and ecotoxicology are available on the market, standardized or not, regulatory or not, that enable us to classify hazardous substances, evaluate potential impacts, and highlight the deterioration or renovation of aquatic environments. Many national or European guidelines exist to assess the effect of substances (e.g., REACH Directive) or environmental quality (e.g., WFD) but without solving the issue. One of the current challenges, despite the growing number of studies, is the development of in situ ecological assessment approaches especially those able to assess the effects of a low level of contaminants. These Responsible editor: Philippe Garrigues

Application of Artificial Neural Networks for Identifying Optimal Groundwater Pumping and Piping Network Layout

The simulation-optimization approach is often used to solve water resource management problem although repeated use of the simulation model enhances the computational load. In this study, Artificial Neural Network (ANN) and Bagged Decision Trees (BDT) models were developed as an approximator for Analytic Element Method (AEM) based groundwater flow model. Developed ANN and BDT models were coupled with Particle Swarm Optimization (PSO) model to solve the well-field management problem. The groundwater flow model was developed for the study area and used to generate the dataset for the training and testing of the ANN & BDT models. These coupled ANN-PSO & BDT-PSO models were employed to find the optimal design and cost of the new well-field system by optimizing discharge & co-ordinate of wells along with the cost effective layout of piping network. The Minimum Spanning Tree (MST) based model was used to find out the optimal piping network layout and checking the hydraulic constraints in the piping network. The results show that the ANN & BDT models are good approximators of AEM model and they can reduce the computational burden significantly although ANN model performs better than BDT model. The results show that the coupling of piping network model with simulation-optimization model is very significant for finding the cost effective and realistic design of the new well-field system.