Skip to content
Search
Type
Lecture

Cost effective analysis of a nuclear facility, London

Event organised by Society for Earthquake and Civil Engineering Dynamics

Date
26 February 2020
Time

This event has now ended

You can catch-up on details or any broadcast and downloads here.

Overview

For a nuclear facility, non-linear seismic analysis is often avoided due to the difficulties in validation. However, in some circumstances it can be appropriate.

The glovebox tunnel system at the nuclear facility in question utilises a chemical bond between the glovebox feet and the supporting concrete slab. The strength of this chemical bond was evaluated via in-situ test work; and was shown to be less than originally assumed in the seismic assessment. This new data had the potential to compromise the containment boundary under seismic loading.

Paul Doyle will explain from the Client’s viewpoint why the testing was undertaken; and how this affected the seismic restraints for a single glovebox, leading on to undertaking some non-linear analysis. He will then explain how this work then led onto a need to consider the behaviour of multiple gloveboxes in a non-linear manner.

The main presentation on the non-linear analysis of this glovebox system will then be given by Steve Horrocks. The methodology for this analysis was to generate an ABAQUS Finite Element model of the coupled concrete building (including soil springs) and the glovebox system. The potential for bond failure was included by using connector elements that had the prescribed bond strength.

Other non-linear stiffness effects were included such as gaps and bolted gasket connections. Non-linear material properties for the stainless and carbon steel were also defined. Modified real seismic time histories were used to drive the model at the soil springs located at the base of the building. The way in which the non-linear behaviour of the model was validated was also considered.

Assessment of the containment boundary was performed at every time step throughout each seismic event. This included monitoring the number of bond failures and the steelwork plastic strain, and calculation of the bolted gasket connection utilisations. The analysis produced a vast volume of data for many hundreds of feet, bolts, gaskets, etc. To make this data understandable, novel methods of graphical presentation was required; and this will be shown in the lecture.

The analyses demonstrated that the containment boundary was not compromised, even though a significant number of bonds failed. Based on this result it was recommended that no costly retrofits were required.

For more information please contact:

Shelly-Ann Russell

Topics: