Shake table testing of conical tank models

A. M.I. Sweedan, A. A. El Damatty

Research output: Chapter in Book/Report/Conference proceedingConference contribution


Conical shaped vessels are widely used as water containment in North America as well as in various places around the world. However, the current standards for the design of water structures, which are usually based on the specifications provided by the American Water Work Association, do not provide a direct seismic design procedure for such structures. This paper includes the first experimental investigation for the dynamic behavior of liquid-filled conical tanks. In this experimental program, a scale model is used to simulate typical steel prototype dimensions. The model, which is made of aluminum, is mounted to a shake table device and subjected to lateral excitation. The dynamic response of the empty as well as the water-filled model is assessed by applying white noise and harmonic excitations at the table level. The liquid sloshing mode as well as the coupled liquid-shell sloshing and impulsive modes are evaluated. The results indicated the significant influence of the geometric imperfections on the dynamic behavior of the tanks. In addition, the results validated the common assumption of de-coupling between the shell vibrations and the liquid sloshing mode.

Original languageEnglish
Title of host publication2000 Annual Conference Abstracts - Canadian Society for Civil Engineering
EditorsD. McTavish, G. Knights, F.M. Bartlett, R.K. Rowe, S. Easa
Number of pages1
Publication statusPublished - 2000
Externally publishedYes
Event2000 Annual Conference - Canadian Society for Civil Engineering - London, Ont., Canada
Duration: Jun 7 2000Jun 10 2000

Publication series

Name2000 Annual Conference Abstracts - Canadian Society for Civil Engineering


Other2000 Annual Conference - Canadian Society for Civil Engineering
CityLondon, Ont.

ASJC Scopus subject areas

  • General Engineering


Dive into the research topics of 'Shake table testing of conical tank models'. Together they form a unique fingerprint.

Cite this