Jalali, N; (2015) Efficient computational techniques for modeling of transient releases following pipeline failures. Doctoral thesis , UCL (University College London).

PDF Navid Jalali Final PhD Thesis-NJ[1].pdf Available under License : See the attached licence file. Download (2MB)

Abstract

This thesis describes the development and extensive testing of a numerical CFD model and a semi-analytical homogenous flow model for simulating the transient outflow following the failure of pressurised pipelines transporting hydrocarbon mixtures. This is important because these pipelines mainly convey highly flammable pressurised and hazardous inventories and their failure can be catastrophic. Therefore an accurate modelling of the discharge rate is of paramount importance to pipeline operators for safety and consequence analysis. The CFD model involves the development of a Pressure-Entropy (P-S) interpolation scheme followed by its coupling with the fluid flow conservation equations using Pressure (P), Entropy (S), Velocity (U) as the primitive variables, herewith termed as the PSUC. The Method of Characteristics along with the Peng Robinson Equation of State are in turn employed for the numerical solution of the conservation equations. The performance of the PSUC is tested against available experimental data as well as hypothetical test cases involving the failure of realistic pipelines containing gas, two-phase and flashing hydrocarbons. In all cases the PSUC predictions are found to produce reasonably good agreement with the published experimental data, remaining in excellent accord with the previously developed but computationally demanding PHU based CFD model predictions employing Pressure (P), Enthalpy (H) and Velocity (U) as the primitive variables. For all the cases presented, PSUC consistently produces significant saving in CPU run-time with average reduction of ca. 84% as compared to the previously developed PHU based CFD model. The development and extensive testing of a semi-analytical Vessel Blowdown Model (VBM) aimed at reducing the computational run-time to negligible levels is presented next. This model, based on approximation of the pipeline as a vessel discharging through an orifice, handles both isolated flows as well as un-isolated flows where the flow in the pipeline is terminated upon puncture failure or at any time thereafter. The range of applicability of the VBM is investigated based on the comparison of its predictions against those obtained using the established but computationally demanding PHU based numerical simulation. The parameters studied to perform testing the applicability of the VBM include the ratio of the puncture to pipe diameter (0.1 – 0.4), initial line pressure (21 bara, 50 bara and 100 bara) and pipeline length (100 m, 1 km and 5 km). The simulation results reveal that the accuracy of the VBM improves with increasing pipeline length and decreasing line pressure and puncture to pipe diameter ratio. Surprisingly the VBM produces closer agreement with the PHU based CFD predictions for two-phase mixtures as compared to permanent gases. This is shown to be a consequence of the depressurisation induced cooling of the bulk fluid which is not accounted for in the VBM. Finally, development and testing of the Un-isolated Vessel Blowdown Model (UVBM), as an extension of the VBM accounting for the impact of initial feed flow and fluid/wall heat transfer during puncture is presented. The performance of the UVBM is tested using a 10 km pipeline following a puncture along its length considering three failure scenarios. These include no initial feed flow, cessation of feed flow upon failure and its termination at any set time thereafter. For the ranges tested, the VBM and UVBM are shown to present considerable promise given their significantly shorter computational run-time compared to the PHU based numerical technique whilst maintaining the same level of accuracy.

Download activity - last month

Export as JSONCSVXML

Download activity - last 12 months

Export as XMLCSVJSON

Downloads by country - last 12 months

Export as CSVXMLJSON

Archive Staff Only

View Item