| dc.contributor |
Deininger, Rolf A. |
|
| dc.creator |
Lee, Byoung Ho |
|
| dc.date |
2016-08-30T16:50:48Z |
|
| dc.date |
2016-08-30T16:50:48Z |
|
| dc.date |
1990 |
|
| dc.date.accessioned |
2022-05-19T13:29:18Z |
|
| dc.date.available |
2022-05-19T13:29:18Z |
|
| dc.identifier |
http://gateway.proquest.com/openurl?url_ver=Z39.88-2004&rft_val_fmt=info:ofi/fmt:kev:mtx:dissertation&res_dat=xri:pqm&rft_dat=xri:pqdiss:9023584 |
|
| dc.identifier |
http://hdl.handle.net/2027.42/128524 |
|
| dc.identifier.uri |
http://localhost:8080/xmlui/handle/CUHPOERS/117239 |
|
| dc.description |
The United States Environmental Protection Agency requires all drinking water authorities to monitor the water quality in their distribution systems to ensure that the water reaching the public will be safe to drink. The regulations prescribe the sampling frequency and the water quality parameters to be monitored. The sampling frequency is based on the size of the population served by the distribution system, and the sampling is to be spaced evenly over time. However, procedures for achieving the goal of representative sampling in a distribution system are not specified. Hence, this research was initiated to develop algorithms and procedures for locating monitoring stations in a water distribution system. A literature search showed that significant changes can occur in the water quality and focused on three major parameters of water quality: chlorine residual, total trihalomethanes, and bacterial concentrations. To select monitoring stations in a rational manner, an understanding of the flow in the distribution system is absolutely necessary. The Kentucky Pipes Model and the WADISO Model were used to predict the flows which were displayed using computer graphics. A computer model was developed to predict the changes of selected water quality parameters for both dynamic and steady-state conditions. An important component of this research is the concept of representativeness, the percentage of population whose tap water is effectively monitored by the sampling stations. Using this concept two algorithms for selecting monitoring stations were developed which maximize the representativeness. The first method uses integer programming to find an optimal set, the second uses a heuristic method to locate the stations using computer graphics. Pathway analyses show the flow from the source of water to a selected sampling station. The methodologies were applied to the water distribution systems in New Haven, Connecticut, and the city of Flint, Michigan. Comparisons between present monitoring stations and the newly developed ones show that the new monitoring stations protect a greater percentage of the population. General graphs were developed which relate the representativeness to the percentage of nodes sampled. These graphs and equations can serve as guidelines for the development of monitoring programs in other networks. |
|
| dc.description |
Ph.D. |
|
| dc.description |
Applied Sciences |
|
| dc.description |
Civil engineering |
|
| dc.description |
Engineering, Sanitary and Municipal |
|
| dc.description |
Environmental science |
|
| dc.description |
Health and Environmental Sciences |
|
| dc.description |
University of Michigan, Horace H. Rackham School of Graduate Studies |
|
| dc.description |
http://deepblue.lib.umich.edu/bitstream/2027.42/128524/2/9023584.pdf |
|
| dc.format |
220 p. |
|
| dc.format |
application/pdf |
|
| dc.language |
English |
|
| dc.language |
EN |
|
| dc.subject |
Distribution |
|
| dc.subject |
Locating |
|
| dc.subject |
Monitoring |
|
| dc.subject |
Networks |
|
| dc.subject |
Stations |
|
| dc.subject |
Water |
|
| dc.title |
Locating monitoring stations in water distribution networks. |
|
| dc.type |
Thesis |
|