No abstract

A Neural Network (NN) modelling approach has been shown to be successful in calculating pseudo steady state time and space dependent Dissolved Oxygen (DO) concentrations in three separate reservoirs with different characteristics using limited number of input variables. The Levenberg-Marquardt algorithm was adopted during training. Pre-processing before training and post processing after simulation steps were the treatments applied to raw data and predictions respectively. Generalisation was improved and over-fitting problems were eliminated: Early stopping method was applied for improving generalisation. The correlation coefficients between neural network estimates and field measurements were as high as 0.98 for two of the reservoirs with experiments that involve double layer neural network structure with 30 neurons within each hidden layer. A simple one layer neural network structure with 11 neurons has yielded comparable and satisfactorily high correlation coefficients for complete data set, and training, validation and test sets of the third reservoir.

No abstract

A model for predicting the concentrations of leachate organics, measured as chemical oxygen demand, in groundwaters below sanitary landfill sites is developed. Simultaneous Michaelis-Menten substrate utilization and microbial mass production equations, with convection and dispersion included for the former, were used for the modeling of biodegradation. For substrate concentrations significantly less than the half utilization rate coefiScient and for microbial populations approaching a steady state the Michaelis-Menten equations are reduced to a first-order reaction kinetic model. The nonlinear model equations are solved using a Galerkin finite element technique, with a Newton Raphson iteration procedure being used to solve the resulting matrix equations. Model sensitivity studies are undertaken to determine the extent to which COD reduction might vary in the field with respect to time and space. The results indicate that substantial removal of leachate organics can be expected within short distances of the landfill even under the conditions of widely varying biological parameters. Model results are compared to the leachate organic reduction occurring below the Canadian Forces Base Borden sanitary landfill. into the adjacent soil media. As a result the potential hazard which these leachates present to groundwater quality adjacent to landfills i• apparent. Despite the strong potential for ground water pollution, comparatively little•research has been done to determine the fate of leachate organic matter in soil and to model its transport in the•groundwater system. As a result the work described in this paper was undertaken to develop a model for organic matter transport in soils adjacent to sanitary landfills and to compare model predictions with field data collected from groundwater zones below a sanitary landfill. ATTENUATION MECHANISMS Mechanisms that reduce organic matter concentrations in groundwater during movement through soil involve both •Now at INTERA Environmental Consultants, Inc., Houston, Texas 77079.,. physicochemical and biochemical processes. The most active physicochemical processes are likely to be dispersion, diffusion, and adsorption, with filtration operative for particulate matter. The biochemical processes involve microbial degradation, either aerobic or anaerobic, with the latter expected to be most active, since oxygen supply at depth within soil-groundwater systems can be expected to be limited. It has been well established that landfill leachates are anaerobically biodegradable [Foree and Reid, 1973; Boyle and Ham, 1974; Pohland, 1975; Chian and DeWalle, 1977a, b; Soyupak et al., 1978]. In addition, information is also available on the kinetics of anaerobic biodegradation of leachates from continuously stirred tank reactors [Foree and Reid, 1973; Pohland, 1975; Soyupak et al., 19•78] and soil column studies [Soyupak, 1979]. Chian and DeWnile [!977a, b] defined four distinct bio-

No abstract

No abstract

Artificial neural network (ANN) methodology has found some recent applications as efficient control tools for satisfying free residual chlorine (FRC) levels at critical locations of water distribution systems. This particular research was started to critically investigate the potential and applicability of the ANN approach as a tool for controlling FRC levels for complex water distribution systems supplied by high quality waters with low chlorine demands. Konyaalti Water Distribution System, operated by Antalya Water and Wastewater Administration, Turkey, has been selected as a pilot. The selected system is complex in structure and supplied with raw water which has high quality and low decay rate of chlorine. The study has shown that ANN models with high predictive power and precision can be developed for such water distribution systems, and that these models can be utilized for forecasting purposes. The data for model building should be collected properly if the developed ANN models are to be utilized as control instruments for FRC levels within water distribution systems.

Porsuk Dam Reservoir (PDR), which is located on Porsuk River, is the main drinking water resource of Eskisehir City-Turkey. Both the river and the reservoir are under the threat of several domestic and industrial point sources and land-based diffuse pollution. The river water quality is very poor with high concentrations of nitrogen and phosphorus compounds at the entrance to Porsuk Reservoir. The reservoir shows symptoms of a hypertrophic lake. The expected responses of the whole river and reservoir system under different pollution control scenarios were estimated to develop plausible water quality management strategies. The adopted scenarios assumed different levels of treatment for the major domestic point sources that include conventional treatment and tertiary treatment. The contemporary Turkish Allowable Discharge Limits (ADLs) and the best available technology choices were the investigated treatment options for the major industries. The expected improvements of water quality characteristics under the management scenario options have been estimated by means of mathematical models. The model choices were the QUAL2E for the river and BATHTUB for the reservoir. Recommendations for different levels of treatment were derived in order to improve the water quality both within the river and in the reservoir.