Marie-Claude Besner

INTRUSION MODELLING AND THE EFFECT OF GROUND WATER CONDITIONS

Intrusion of contaminants into water distribution systems due to negative pressures is a complex phenomena that has been theorised and is an active area of research. Work has identified the existence of contaminants in soil and ground water surrounding pipes, and has investigated modelling the risk to human health should contaminants enter water distribution systems. However, there is a lack of understanding regarding the critical interaction between the pipe, the leak aperture and the surrounding ground and water. Typical intrusion models assume a simple orifice relationship, with inflow volumes proportional to the square root of the difference between the pipe pressure and the external hydrostatic pressure. This is shown here through computational modelling to be an overly simple relationship for leak behaviour that does not take into account the existence or properties of a porous media external to the pipe. In this paper the authors will discuss the construction of computational fluid dynamics (CFD) models of the intrusion process due to transient events and describe results that suggest the simple orifice equation is not a sufficient model. In the CFD calculations the surrounding ground water is modelled as a saturated porous media and it is shown that the properties of the media and the leak geometry have a large effect on the relationship between the pressures and flow rate. It is also shown that the risk of intrusion should be considered from contaminants that originate from both above and below the depth of the pipe. Further, the idea of a zone of influence surrounding the leak point, due to the leak size and the magnitude and duration of the negative pressure, is introduced as a possible measure of the level of risk of the intrusion event.

Negative Pressure Events in Water Distribution Systems: Public Health Risk Assessment Based on Transient Analysis Outputs

Transient analysis of a pump trip was conducted on a full-scale distribution system (DS) equipped with high-speed pressure transient data loggers at the outlet of the water treatment plant (WTP) and at 12 DS sites. Following the calibration of the transient model (~16,000 nodes) with transient pressure recordings, intrusion volume computations were performed considering two intrusion pathways: leakage orifices and submerged air vacuum valves (AVVs). As expected, the estimated intrusion volumes through submerged AVVs are considerably larger than those through leakage orifices. Water quality modeling was conducted in order to evaluate the spatiotemporal dispersion of the intruded water, assumed to be contaminated with Cryptosporidium oocysts. A point estimate of the maximum probability of infection was then computed at each DS node using the negative exponential model. The estimated maximum probabilities of infection were displayed on the DS map and the model assumptions are discussed. This exercise has underlined important risk gradients and the localized occurrence of very high probabilities of infection, suggesting that a global risk analysis might be misleading. This project is the first attempt at quantifying public health risks induced from low pressure events in a large scale system (supplying ~400,000 people) based on actual negative pressure recordings.

Cryptosporidium and Giardia in Wastewater and Surface Water Environments

and spp. are significant contributors to the global waterborne disease burden. Waterways used as sources of drinking water and for recreational activity can become contaminated through the introduction of fecal materials derived from humans and animals. Multiple studies have reported the occurence or concentrations of these pathogens in the environment. However, this information has not been comprehensively reviewed. Quantitative microbial risk assessment (QMRA) for and can be beneficial, but it often relies on the concentrations in environmental sources reported from the literature. A thorough literature review was conducted to develop an inventory of reported and concentrations in wastewater and surface water available in the literature. This information can be used to develop QMRA inputs. and (oo)cyst concentrations in untreated wastewater were up to 60,000 oocysts L and 100,000 cysts L, respectively. The maximum reported concentrations for and in surface water were 8400 oocysts L and 1000 cysts L, respectively. A summary of the factors for interpretation of concentration information including common quantification methods, survival and persistence, biofilm interactions, genotyping, and treatment removal is provided in this review. This information can help in identifying assumptions implicit in various QMRA parameters, thus providing the context and rationale to guide model formulation and application. Additionally, it can provide valuable information for water quality practitioners striving to meet the recreational water quality or treatment criteria. The goal is for the information provided in the current review to aid in developing source water protection and monitoring strategies that will minimize public health risks.

QUANTIFICATION OF THE RELATIVE IMPORTANCE OF FACTORS CONTRIBUTING TO INTRUSION IN A DISTRIBUTION SYSTEM USING A FULL FACTORIAL DESIGN

Transient analysis of a pump trip event was conducted on a full-scale distribution system (~16,000 nodes) equipped with high speed pressure transient data loggers at the outlet of the water treatment plant (WTP) and at several distribution system (DS) sites. A calibrated transient model was used to perform intrusion volume computations considering two possible intrusion pathways: leakage orifices and submerged air vacuum valves (AVVs). For this pump trip event, a full 3-level 4-factor factorial design (3 4-0 ) with 82 runs was completed to understand the relative impact of the following factors on the total intrusion volume: (1) the external head of untreated water on leakage orifices, (2) the external head of untreated water on the outlet orifice of submerged AVVs, (3) the leakage rate (which translates into orifice size), and, (4) the diameter of AVVs’ outlet orifice. Although some of these factors may be highly uncertain, fieldwork (including installation of piezometers and visits of air valve vaults) is currently being conducted, and realistic values were assigned to these factors. The analysis showed that the factors and interactions associated with AVVs had a significant effect on the total intrusion volume. The importance of the external head on AVVs’ orifice on the total intrusion volume was highlighted in a plot of marginal means. When intrusion through both pathways occurs concurrently, the interactions between these intrusion flows influence the total intrusion volume. The latter can even decrease when the external head on leakage orifices increases creating surge dampening effects large enough to significantly reduce intrusion through submerged AVVs.