Tawalo, Ali (2025) Strategies for Safeguarding Natural Gas Pipeline Network Systems Against Natural Hazards. [Tesi di dottorato]

Documento PDF Strategies_for_Safeguarding_Natural_Gas_Pipeline_Network_Systems_Against_Natural_Hazards.pdf Visibile a [TBR] Amministratori dell'archivio Download (15MB) | Richiedi una copia

Abstract

Natural gas pipeline network systems are an indispensable type of infrastructure because of their vital role in transporting essential energy resources. Given their widespread nature, pipeline network systems may cross unstable slopes, thereby exposing them to potential deformations caused by slope movements. This research develops strategies to safeguard natural gas pipeline network systems exposed to landslide displacements caused by rainfall infiltration and earthquakes. It focuses on the buried pipeline components, recognizing that enhancing their resilience against these natural hazards is crucial for improving the safety of the entire pipeline network system. For this purpose, this research proposes a numerical framework for assessing the risk of buried steel natural gas pipelines exposed to rainfall-induced landslide displacements, along with risk-based inspection and maintenance strategies, as well as mitigation measures associated with the pipeline. This numerical framework starts with deterministic risk assessment by assessing the risk of buried steel natural gas pipelines caused by specific rainfall scenario-induced landslide displacements. It is then improved to account for the uncertainty associated with the assessment of rainfall-induced landslide displacements and to incorporate the probabilistic hazard of seismic-induced landslide displacements. Moreover, this research proposes a simplified risk assessment framework for pipelines exposed to landslide displacements induced by groundwater table fluctuations. This framework can replace the numerical framework, providing simplified risk based inspection and maintenance strategies. In detail, the research employs advanced numerical modeling to analyze slope-atmosphere interaction. Using the Miscano landslide in Southern Italy as a case study, coupled flow deformation analyses are performed in Plaxis 2D employing both the Mohr-Coulomb and Hardening Soil with small-strain stiffness (HSsmall) constitutive models to simulate soil behavior.Calibration and validation against two years of field monitoring data demonstrate that both models can reliably predict pore pressure and associated yearly surface displacements, while HSsmall offers a more realistic representation of cumulative displacements compared to Mohr-Coulomb. The choice of initial conditions proves critical, as the analyses starting from the initial phase, rather than the calibration year, more accurately reproduce displacement trends during the validation year when the aforementioned constitutive models are used. The combined effect of rainfall and seismic loading on the landslide behavior is also investigated by subjecting the Miscano numerical model to different rainfall scenarios, obtained from meteorological stations close to the site of Miscano, and seismic records compatible, on average, with the site’s design spectrum obtained using REXEL tool. Results highlight that hydrological conditions, particularly the position of the groundwater table resulting from rainfall infiltration, significantly influence the seismic contribution to landslide displacements. Moreover, seismic induced displacements are primarily governed by the shear strength and depth of the sliding surface. The research develops a numerical framework for assessing the risk of buried steel natural gas pipelines exposed to axial landslide displacements induced by rainfall. The developed framework combines the aforementioned 2D numerical model of Miscano, with a 3D soil-pipe interaction numerical model implemented in the finite element code Plaxis 3D. The 2D numerical model of Miscano is used to evaluate the hazard by assessing the landslide displacements induced by a specific rainfall scenario along the slope. While the vulnerability of buried steel pipelines is assessed using fragility functions that describe the probability of pipeline failure, where local buckling is assumed to be the potential structural instability. These fragility functions are derived from statistical analyses of numerical data based on 3D soil-pipeline interaction simulations, correlating the differential permanent landslide displacements with the pipeline strains. The strains are derived from a parametric study that accounts for variations in pipeline dimensions, backfill material properties, steel grades, and internal gas pressure. The findings show that larger pipeline dimensions and higher steel grades reduce vulnerability, while increased backfill compaction increases it. Event Tree Analysis is used in the context of this framework for assessing the potential consequences of pipeline failure. It reveals that a flash fire from delayed local ignition, following local buckling failure, has the highest probability among potential failure consequences. The framework is further extended to include risk-based inspection and maintenance strategies, ensuring that monitoring and intervention efforts can be optimized according to the assessed risk. The research enhances the previous numerical framework by improving the hazard assessment part, considering the uncertainty associated with assessing both rainfall-induced and seismic-induced landslide displacements through probabilistic approaches. This uncertainty arises from two main sources: the first one is related to the displacements themselves, which stems from limited knowledge of soil properties and the slope’s response to triggering events (rainfall or earthquake), and the second one is associated with natural randomness in the intensity of triggering events over time. The developed probabilistic approach for each natural hazard consists of three main steps: sensitivity analysis, development of the probabilistic approach, and applying the total probability theorem. For rainfall-induced landslide displacements, the sensitivity analysis demonstrates that permeability, rainfall patterns, groundwater depth, and distance from the landslide toe strongly influence displacement variability. Uncertainty in landslide displacements is addressed by performing statistical analysis of numerical data, in terms of horizontal displacements, obtained from hundreds of coupled flow-deformation analyses under varying hydro-mechanical parameters, based on their probability distributions, and multiple rainfall scenarios. Moreover, uncertainty in rainfall-induced landslide displacements can be successfully modelled using a log-normal distribution. Rainfall variability is similarly modelled using a log-normal distribution fitted to long-term meteorological data obtained from a meteorological station close to the Miscano location. Using total probability theorem, the approach combines these two uncertainty sources to establish hazard curves relating displacement to exceedance probability over one year. For seismic-induced landslide displacements, a similar probabilistic approach is developed. Sensitivity analysis demonstrates that Peak Ground Velocity exerts greater control over displacements than Peak Ground Acceleration or Arias Intensity, and that shear strength of the sliding surface dominates the slope response. Uncertainty in displacements is again modeled as log-normal through statistical analysis of numerical data, in terms of horizontal displacements, obtained from hundreds of dynamic simulations. At the same time, the temporal variability of PGV is incorporated using a Poisson distribution. This probabilistic approach yields hazard curves, using the total probability theorem, that can be directly integrated into the pipeline risk assessment framework developed previously. Alongside the detailed numerical framework with its improvements mentioned previously, a simplified probabilistic approach is proposed to support performance based risk assessment of buried natural gas pipelines subjected to landslide displacements driven by groundwater table fluctuations. In its hazard assessment component, this approach extends an existing deterministic analytical model into a probabilistic one by incorporating the spatial variability of landslide displacements, using a semi-variogram tool, and the temporal variability of groundwater fluctuations through the annual exceedance rates. Pipeline vulnerability in this framework is quantified through empirical fragility functions using empirical data that relates permanent ground deformation to repair rates. The framework is applied to the Vallcebre landslide as a case study, with risk quantified as the probability of having at least one break or leak per 100 m length of pipeline. Nevertheless, the developed empirical fragility functions present limitations, particularly due to the lack of detail in the underlying empirical data regarding the severity and cost of required repairs and the orientation of permanent ground displacement. In addition, the analytical model adopted in the hazard assessment introduces further constraints. Given these limitations, the proposed framework can be considered as a preliminary tool for risk assessment.

Downloads

Actions (login required)

Modifica documento