TY - BOOK

T1 - Numerical modeling of pipe leakage in variably saturated soil

AU - Peche, Aaron

PY - 2019

Y1 - 2019

N2 - Pipe leakage related to defect urban sewer and stormwater pipe networks may lead to subsurface contamination, reduction of groundwater recharge or a significant decrease of the groundwater table. The quantification of pipe leakage is challenging, mostly due to the uncertain forming of a colmation layer in the defect vicinity and inaccessibility of both, pipe defects and the surrounding soil. Numerical models can be used to quantify leakage. At present times, only few physically-based pipe leakage models exist, all of which either neglect or simplify variably-saturated flow. In the present dissertation, a novel and unique three-dimensional physically-based pipe leakage model for variably saturated soil is presented. The model consists of the newly implemented coupling between the pipe flow simulator HYSTEM-EXTRAN and the unsaturated-saturated flow simulator OpenGeoSys. The coupling is based on updating of boundary conditions and source terms. The interprocess data transfer is realized using a shared-memory. The pipe leakage model is successfully validated and verified using a newly generated benchmark library for pipe leakage models. Benchmarks are based on two physical experiments described in literature and two newly derived analytical solutions. A novel method for upscaling pipe leakage is presented. The method enables to significantly reduce the local refinement of spatial discretization in the pipe vicinity, which leads to a substantial reduction of computational costs. The method is based on leakage functions. Two leakage functions representing both, sewer and stormwater pipe leakage are presented. Accuracy and time efficiency of the upscaling method is demonstrated by comparing results from a fully discretized model and an upscaled model. In the present dissertation, the pipe leakage model is applied to several case studies to investigate the pipe leakage process. Model setups represent (i) a single defect, (ii) a leaky sewer pipe of 30 m length, and (iii) a 53 km long defect stormwater pipe network in an urban catchment. Results of the single defect and leaky sewer pipe model (models i,ii) show that leaky pipes can hydraulically disconnect from groundwater. It is found that, for a given pipe water level, pipe water exfiltration converges as the groundwater table is lowered. Further, it is found that pipe water exfiltration increases as the intensity and duration of pipe flow events increase. The temporal distribution of pipe flow has a negligible effect on pipe water exfiltration. Results of the model representing a defect stormwater pipe network in an urban catchment (model iii) show the impact of pipe leakage on urban groundwater. It is shown that groundwater infiltration into a largely defect pipe network may be in the order of annual groundwater recharge. Further, it is shown that groundwater infiltration may reduce the groundwater table by several meters and that the groundwater table may be lowered to the elevation of the pipe network.

AB - Pipe leakage related to defect urban sewer and stormwater pipe networks may lead to subsurface contamination, reduction of groundwater recharge or a significant decrease of the groundwater table. The quantification of pipe leakage is challenging, mostly due to the uncertain forming of a colmation layer in the defect vicinity and inaccessibility of both, pipe defects and the surrounding soil. Numerical models can be used to quantify leakage. At present times, only few physically-based pipe leakage models exist, all of which either neglect or simplify variably-saturated flow. In the present dissertation, a novel and unique three-dimensional physically-based pipe leakage model for variably saturated soil is presented. The model consists of the newly implemented coupling between the pipe flow simulator HYSTEM-EXTRAN and the unsaturated-saturated flow simulator OpenGeoSys. The coupling is based on updating of boundary conditions and source terms. The interprocess data transfer is realized using a shared-memory. The pipe leakage model is successfully validated and verified using a newly generated benchmark library for pipe leakage models. Benchmarks are based on two physical experiments described in literature and two newly derived analytical solutions. A novel method for upscaling pipe leakage is presented. The method enables to significantly reduce the local refinement of spatial discretization in the pipe vicinity, which leads to a substantial reduction of computational costs. The method is based on leakage functions. Two leakage functions representing both, sewer and stormwater pipe leakage are presented. Accuracy and time efficiency of the upscaling method is demonstrated by comparing results from a fully discretized model and an upscaled model. In the present dissertation, the pipe leakage model is applied to several case studies to investigate the pipe leakage process. Model setups represent (i) a single defect, (ii) a leaky sewer pipe of 30 m length, and (iii) a 53 km long defect stormwater pipe network in an urban catchment. Results of the single defect and leaky sewer pipe model (models i,ii) show that leaky pipes can hydraulically disconnect from groundwater. It is found that, for a given pipe water level, pipe water exfiltration converges as the groundwater table is lowered. Further, it is found that pipe water exfiltration increases as the intensity and duration of pipe flow events increase. The temporal distribution of pipe flow has a negligible effect on pipe water exfiltration. Results of the model representing a defect stormwater pipe network in an urban catchment (model iii) show the impact of pipe leakage on urban groundwater. It is shown that groundwater infiltration into a largely defect pipe network may be in the order of annual groundwater recharge. Further, it is shown that groundwater infiltration may reduce the groundwater table by several meters and that the groundwater table may be lowered to the elevation of the pipe network.

U2 - 10.15488/4956

DO - 10.15488/4956

M3 - Doctoral thesis

CY - Hannover

ER -