Martin BrownGE19 Bridge Displacement

Director, Health, Safety and Environment

London Rail

Outline

• Some Background• GE 19 and the incident • Why it occurred• Lessons for Design, Construction and

Constructability and Site monitoring• Some conclusions and suggested next steps

East London Railway Extension

The East London Line Extension, connects the old London Underground East London Line (running from New Cross to Shoreditch, including Marc Brunel’s Thames Tunnel), part of the existing suburban line to Crystal Palace and Croydon, and some new Infrastructure. Phase 2 is a proposal to link to Clapham Junction

GE19

Bridge GE19

• Bridge set at 1:30 gradient, falling from west to east

• Situated near the Broadgate Tower, near to the Liverpool Street Station throat

Bridge GE19 : Plan View

• Set at 60 degrees to the Great Eastern Main Line, which is a six track • Lying almost due West to East

GE19 in numbers

• Spans 84 metres, with 8 metres deep truss• Made from steel weighing 812 tonnes• Bridge decked out with 417 Omnia planks,

each weighing approximately 100k • Approximately 500 tonnes of concrete was

cast as counterweight (45% of total length) pre-launch, at the East end

• The bridge sits on four bearings

What Occurred• Launch successfully completed during a possession

weekend 3 – 5th May 2008, landing within 1 mm of design• Between 5th and 21st May preparation and lowering work

was undertaken• Mid-May it was realised the bridge had moved, and needed

to be plan jacked approximately 38 mm (East to West), against lateral restraints

• Thermal expansion was about 20mm during day• With the plan jack completed, final lowering was scheduled

to be completed on 29th – 30th May• On 28th May, at 19.15, Bridge dropped 200mm at East End

onto the permanent bearing

Bearings Layout• At launch, the base of the

bridge was 650mm above the East bearing, and 400mm above the West

• The bridge sat on steel plates on top of the final bearings

• The bearing set up, provides constrained movement as shown. The West end is fixed

• The bearing have transport restraints

• The bearings have smooth finish stainless steel contact surfaces

Movement is fixed

Lateral movement North-South

Longtitudinal movement East-West

Free movement

500 tonne Hydraulic Jack,

with security collars

Top cap of abutment

Permanent bearing

Lower Member of Bridge GE19

Temporary spacer plates

Temporary PTFE

Covered plates

3 degree temporary

wedge

Note the inclination of the bridge has been enhanced for clarity

Some load assumptions

• The load on the eastern end bearings was about 450 tonnes• With the slope about 3.5 degrees, simple assumptions give a

resolved horizontal down slope load of about 18 tonnes• Again with some assumptions, the breakout friction at the PTFE

plate would be 3% to 4%, giving a stable position over some time • Cooling temperature change could lead to movement of 2 to 3 mm,

converting static to dynamic levels• Dynamic friction would be 1% or less• The transit bolts would have provided maybe 2 tonnes each• However once movement began, the spacers would begin to be

squeezed out, and the transit bolts bent. This is what we think happened

Omnia Planks

Note this plank shows signs of damage from the bridge displacement

Note the lacing in both directions tosecure the planks in position,

fitted post incident

Omnia planks issues• The Omnia planks were held in place by ‘contraband’

tape, with a filler material in the gaps between planks• Shear pins were fitted between the planks, to take up the

loads when concreted• The planks were not laced, following a value engineering

and pre-launch review. There was no secondary means of preventing the planks dropping onto the live railway

• When the bridge was displaced the planks were subject to a significant movement in some places; 5 fell onto the track, and several others were dislodged

• The method was based on previous successful installations

Bridge drops (causing disruption of service)

PTFE pad inserted in the wrong position

Inadequate understanding of risks

by supervision

Unsecured Omnia planks falls onto track

Risk assessment discounted possibility

of planks falling

Risks not adequately identified and analysed in assessment

MS short time change procedure not

followed

Design of temporary works does not

consider realignment during jack down

Unexpected movement down

slope dislodges plates

Initiating Event

Immediate cause

Contributory Causes

Root Causes

GE19 Bridge Incident – Root cause map

No secondary means of securing planks

Temporary support bearings fail

Unsecured planks dislodged by

movement of Bridge

Design did not fully encompass the

practicability of the temporary works for lowering the bridge

Cumulative design checks reinforce expected success of

planned works, hence reducing perceived risks

Inadequate dynamic risk assessment

undertaken

Wrong instruction given

No written instructions in place or suitable

briefing

Method statement requiring interlacing

was discounted

Inadequate controls over change of

methods

Deviation from method

Statement

No evidence of explicit standards and procedures in place for management of temporary works

Once launch complete there was a reduced focus on the lowering and temporary works

controls

Lessons from the Displacementof the Bridge

• Design of the Temporary Works• Lack of explicit standard or work instruction for

placement of the PTFE• Reduced focus, once launch was complete• Design of bridge did not fully encompass

constructability of all elements• Failure to understand the high consequence/ low

probability risks involved

High load sliding elastomeric bearings have been specified for concrete box girders.Preformed fabric is bonded to a steel plate that is recessed to receive a PTFE sliding surface.The top is stainless steel on a steel backing plate. Anchorage to the upper and lower concretesurfaces is with end welded studs. Under these conditions, the following controls arerecommended:

♦ The preformed fabric should be 2.0 in. thick. ♦ The average bearing pressure on the preformed fabric due to dead load and live

loadwithout impact should not exceed 2,000 psi.

♦ The average bearing pressure on the PTFE surface should not exceed 3,000 psi. ♦ Top and bottom surfaces should be level. Concrete or epoxy grout should be cast

against the in-place bearing.

Bridge Design ManualDecember 2001© by Texas Department of Transportation(512)

Conclusions

• These circumstances were (as far as we can tell) unique• If our analysis is right, the potential for failure had not

been fully considered before, why was this ? • Once explained, everyone felt the risks were obvious. If

our analysis is wrong what did happen ?• The management of the design, in parcels of work, may

have assisted in the lack of design integration• Greater use should be made to ‘what if’ analysis where

high potential risk exists, especially where there can be further consequential risk

Next Steps• If we are right, we need to change the guidance on how,

where (and when) PTFE plates are used in structures

• For those who ‘sit behind the desk’ planning and designing, we need to ensure production of the information for ‘those on the tools’ to act correctly, this is an educational issue

• We need to reinforce an approach that questions design at a fundamental level, looking at the ‘what if’. This is also an educational issue

• We need to understand more how Human Factors impact on undertaking fundamental and proactive analysis, design checking and site inspection, this may be a research issue

Questions to :martinbrown@tfl.gov.uk

Full Report can be found at :

http://www.tfl.gov.uk/corporate/projectsandschemes/networkandservices/eastlondonrailway/2111.aspx

orhttp://www.tfl.gov.uk/assets/downloads/GE19-Incident-Main-

Report.pdf