THE QUALITY PROBLEM IN WATER DISTRIBUTION SYSTEM

In the last decades attention towards human health and the rise of life quality standards have lead to a higher interest in the quality of water flowing within pressurized flow systems, such as Water Supply Systems (WSSs) and urban Water Distribution Systems (WDSs). Indeed, it is even more necessary to guarantee high quality levels for drinkable water during the ordinary operation of the system, trying to avoid, for example, the formation of harmful elements such as disinfection by-products (DBPs). Another important issue is the intentional contamination, through the introduction of biological, chemical or radio-active contaminants, whose effects could have a strong repercussion on both public health and safety feeling. As a consequence, to adequately simulate contamination phenomena and individuate alternative intervention policies, such as optimum sites for chlorination booster stations (very important, in ordinary conditions, to decrease DBPs) and monitoring early warning stations (very important to deal with emergency conditions rising from intentional contamination), it is necessary to use appropriate analysis tools and therefore produce specific mathematical models in order to analyze space and time variability of the parameters that characterize the quality of water flowing within both WDSs and WSSs. One of the most complete models quoted in literature for the analysis of changes of water quality characteristics along WDSs, is undeniably the one proposed by Rossman et al. [10], which allows to evaluate the variability of chlorine concentration in the water distribution systems. The model uses a Lagrangian approach and is based on the subdivision of the water volume contained in each pipe into a given number of segments (elementary cells), which have the same velocity inside the pipe as that of the flow and behave like completely stirred tanks. Because of its characteristics, last version of this model, denominated EPANET 2 [10], is considered, at present, the world standard, even though it presents some problems, hereafter described. Starting from these considerations, the aim of this paper is to propose a new mathematical problem, denominated QualSim_PFN (Quality Simulation in Pressurized Flow Networks), which is able to overcome EPANET's lacks. As a consequence, in this paper it will first described the EPANET model, then the approach used in the QualSim_PFN model and, finally, it will be synthesized the results of a few runs carried out in order to compare their features and performances. In particular, the contaminants moving within WDSs are usually characterized by very low concentrations and small particle size. As a consequence, they can be considered completely dissolved in the water and, thus, unable to change the flow characteristics (such as velocity and pressure fields). For these reasons, the analysis of time and space changes of quality parameters of water flowing within WDSs is usually carried out in two steps. In the first step, using appropriate hydraulic simulator (hydraulic module), the main characteristics of flow (flow velocities in the pipes, pressure heads at nodes, water levels at reservoirs, etc.) are determined (eventually, for each operating conditions if a period of time characterized by different users’ demands and/or different working conditions of pumps, valves, pipes, etc., have to be considered). In the second step, starting from the knowledge of hydraulic variables (in particular: i) the mean flow velocities along the WDSs’ sides; ii) the water volume in the reservoirs; and iii) the discharge delivered at nodes or in-flowing/out-flowing in/ from the reservoirs), assigned the boundary conditions concerning the contaminants input at nodes and/or reservoirs and, eventually, the coefficients of reaction kinetics, using appropriate quality changes simulator (quality module), the changes in time and space of contaminant(s) concentration(s) are evaluated, together with the mass of contaminant(s) delivered at each node in each time interval and the time spent from a given contaminant input to arrive at various WDSs’ nodes (Early Warning Time). As a consequence, the changes of water quality characteristics within a WDS are strictly linked to the functioning of the distribution system itself. For this reason, before deepen into the ‘quality’ issue, it is important to focus on the WDS ‘hydraulic’. Generally speaking, the module directed to the analysis of the hydraulic behaviour of the water system should be able to simulate different operating conditions that can occur in the system, such as steady and unsteady flows. In order to determine the flow parameters, assuming as certain the geometrical and hydraulic characteristics of the system, initial water levels in the tanks and reservoirs, and the users’ water demand in each time interval in which it is possible to subdivide the whole operation period to examine, it is necessary to solve a set of non linear equations formed by the continuity equations at nodes, the loop equations for each closed circuit existing in the system and roughness formulas for each side. In its turn, the analysis of contaminants propagation should be based on a module directed to the resolutions of a set of continuity equations for contaminant(s), written as for the WDSs’ nodes, as for each of the cells in which every side will be subdivided, as for each tank and reservoirs. These conservation mass equations have to consider different aspects: the mass already present and/or introduced into the tanks and reservoirs, the mass of the substances out-flowing at the water consumption nodes, the mass of the substances that are present and pass in the each side of the WDS as well the kinetic reactions of chemical components, the decay, volatilization phenomena in the tanks/ reservoirs, etc.