This thesis presents a comprehensive investigation of unsteady friction modeling. The research falls into two areas: assess the necessity of unsteady friction modeling and improve on existing models.

The importance of unsteady friction in two pipe systems connected in series is determined. A new index (r_e) measuring the importance by considering the energy dissipation rate is proposed and tested. It shows that this index is capable of estimating the importance in a fair manner and can be easily obtained. Numerical experiments of systematically designed transient cases are performed. The energy behaviors show that the importance of unsteady friction for a two pipe system is always larger than that for its corresponding single pipe systems when the diameter difference of the two...[ Read more ]

This thesis presents a comprehensive investigation of unsteady friction modeling. The research falls into two areas: assess the necessity of unsteady friction modeling and improve on existing models.

The importance of unsteady friction in two pipe systems connected in series is determined. A new index (r_e) measuring the importance by considering the energy dissipation rate is proposed and tested. It shows that this index is capable of estimating the importance in a fair manner and can be easily obtained. Numerical experiments of systematically designed transient cases are performed. The energy behaviors show that the importance of unsteady friction for a two pipe system is always larger than that for its corresponding single pipe systems when the diameter difference of the two pipes is relatively small; when the diameter difference is large, the longer pipe in the system governs the importance of unsteady friction.

Improvements in “frozen” viscosity assumption of weighting function based model (WFBM) are made. “Frozen” viscosity means that the magnitude and distribution of the turbulent viscosity are fixed to its pre-transient state. The comparison between numerical results and measured pressure head traces from laboratory and field tests show that this assumption is valid during the early stages of the transient but its validity reduces with the duration of the transient. To correct this deficiency, a modification to the WFBM is proposed, in which the effects of time dependency is considered. This modification shows a good match with the analytical and field measured pressure wave amplitude.

The adaptability of relaxation time-based model (RTBM) is improved. RTBM is physically based and highly efficient in determining unsteady friction. The major limitation of existing RTBM is the pre-calibration of empirical coefficients before applying to different pipelines. A new RTBM model is proposed with the expressions of the empirical coefficients obtained from curve fitting to numerical results of WFBM. Verification by other numerical models and experimental data shows that the new model performs well in different pipelines without any pre-calibration.