MANAGEMENT AND DESIGN

The Thames Barrier

R.W. Homer, OBE, TD, BSc, CEng, FICE, MIMechE

Indexing terms: Engineering administration and management, Project and production engineering, Project management, Management

Abstract: An area of over 100 km2 of industrialand residential property was at increasing riskfrom flooding, including some 700,000 residentsand the central section of the undergroundrailway. Attention was drawn to this risk in a1966 report by Sir Herman Bondi, and the GLCwas empowered by the Thames Barrier Act of1972 to construct a barrier at Woolwich Reach,13 km downstream from London Bridge. Thealternatives of a barrier and a barrage were exam-ined, and the final design employs rising sectorgates, providing four main navigation openingeach of 61 m, and a further six of 31 m. Rotationof the gates is by radius arms, driven from hydrau-lic cylinders, with three 140 kW pump powerpacks on each of the principal piers. The site isprovided with three independent sources of elec-tric power, connected to the piers via duplicatedaccess tunnels. For the construction there werethree main and 20 other contractors, and sitework began in 1975, erection of the gates in 1980,and the barrier was first completely closed in1982. The projects passed through three stages:feasibility, detailed design and contract letting,and construction and commissioning. The projectwas controlled through monthly site meetings.Total cost escalated from £110 x 106 at 1973prices to £440 x 106 on completion in 1984, whichhas to be set against the estimate of £5000 x 106

as the cost of a major flood disaster.

1 IntroductionAs a consequence of the development of London as amajor seaport and the capital city of the UnitedKingdom, some 110 km2 of densely built-up industrialand residential area have a ground level below that of thehighest water level reached in the River Thames whichflows through the centre of the city and is tidal through-out the whole of the area.

The flood risk arises as a result of storm surges in theestuary, caused by northerly gales in the southern NorthSea which have raised water levels on the rising tide byas much as 4 m and actual high water by 2.6 m. Normalastronomical tides rise to within 1.4 m of the top of theflood defences in the city, so that a 2.6 m surge wouldcause catastrophic flooding, and no guarantee can begiven that a 4.0 m surge could not occur at high water, orthat even higher surges could develop in the future.

In addition, the record of surge tides over the past 200years shows that they are rising ever higher as the years

Paper 5572A (M2) received 17th February 1986The author is at Heortnesse, Hollybush Hill, Stoke Poges, Slough SL24PX, United Kingdom

752

pass (Fig. 1). The main cause of the adverse change ofsome 0.75 m per century is thought to be the general risein sea levels due to the melting of the polar ice caps, aug-mented in Southern England by a tilt of the country

new flood defence level 7.20m

x>o 5

interim defence level 1971/1972 5.80m

1930 defence level 5.28 m

i rr. .T.I.o o o o o oOO O CM -J I P OOt^ oo ao 00 CO ao

O O O O O O O O OO CM ^J CO 00 O CM vj 10en en en en en o o o o" — • — • — •— •— C M C M C M C M

years

Fig. 1 Floodtide levels at London Bridge

about an axis from the Bristol Channel to the Humber.This tilt is considered to be a reaction to the melting ofthe vast mass of ice which existed over the north ofEngland and Scotland during the last ice age some 20000years ago, the tectonic plate on which the country is sitedtilting on the underlying plastic substratum. A number ofother factors may contribute to the phenomenon [1]. Inthe area of the city at risk of tidal flooding some 700 000people live, the number increasing during the workingday to over 1 million; vast investment has been made inhousing, industrial premises and infrastructure. The vitalcentral section of the underground railway system wouldbe flooded, some 70 stations and 74 km of the trackbeing affected. It has been estimated that the cost of amajor flood in the area could approach £5000 x 106 (Fig.2).

In a report to the UK Government in 1966, ProfessorBondi (now Sir Herman Bondi) concluded that such a

1EE PROCEEDINGS, Vol. 134, Pt. A, No. 9, NOVEMBER 1987

catastrophe was a risk no responsible government couldaccept if action could be taken to prevent it [2]. As aresult, the Government asked the Greater London

the standard to be adopted was largely a matter of judg-ment, although benefit cost studies were carried out todemonstrate that the recommended standard was reason-

improvedflood defences

Fig. 2 Map showing potential flood areas before completion of the barrier

Council to carry out a full investigation into the problem,and the decision was subsequently taken to implement arecommendation, made by the Waverley Committeereporting on the East Coast floods of 1953, that a tidalflood barrier be constructed in the Thames Estuary [3].The Thames Barrier Act which empowered the GreaterLondon Council (GLC) to construct such a flood barrierat the Silvertown site in Woolwich Reach, 13 km down-stream of London Bridge, received the Royal Assent inAugust 1972, and work on site commenced early in 1975.

2 Design of the Barrier

Prior to the GLC investigation, a number of proposalshad been made for structures to be built in Long Reachor Crayfordness, but owing to the onerous navigationalrequirements and the large area of the waterway at thesepoints, the structures proposed were both complex andcostly, and for this reason, no decision was taken to buildany of them.

The GLC investigation tackled the problem from firstprinciples, setting up working parties to examine the rele-vant factors such as surge-tide phenomena, navigationalrequirements etc. A considerable number of sites andboth barrier and barrage schemes were examined. Theterm barrier is used to describe a structure across awaterway that is normally open, allowing a free flow oftide through the structure, but which has gates that canbe closed when a potentially dangerous surge tidedevelops, preventing the large volume of water frompassing upriver and causing flooding. In contrast, abarrage is a structure which permits water to flow in adownriver dirction only, controlling water flow contin-uously whether surge-tide conditions develop or not.

A major factor in determining the final scheme was thestandard of protection to be provided for London andthe built-up areas in the lower estuary. The decision on

ably economic. At an early stage, it was accepted that,both as a matter of equity and for practical reasons tomitigate opposition to the selected scheme, the samestandard of flood defence would be provided for all built-up areas, whether upriver or downriver of any barrier orbarrage structure.

The standard selected was that of containment of ahigh water in the estuary, liable to occur on a one in athousand chance in any particular year. In addition,allowance was made for the apparent adverse trend ofhigh waters up to the year 2030. In the light of presentknowledge, some 15 years on, the rate of rise of highwaters was probably overestimated, and both barrier anddownstream flood defences are capable of containing theone in a thousand tide well into the second half of thenext century.

The schemes considered ranged from a barrage in theouter estuary from Margate to Clacton, on the lines ofthe Eastern Scheldte scheme now being built, to raisingflood defences along the whole length of the estuary tothe tidal limit. The final choice as already mentioned, wasto construct a flood defence barrier in the western half ofWoolwich Reach and to raise flood defences downriverto the outer estuary.

Rendel, Palmer and Tritton (RPT) were appointedconsulting engineers for the selected scheme by the GLC,and work was put in hand to complete the final designand to proceed with the construction.

The design requirements laid down for the guidance ofthe consulting engineers were as follows:

(a) four main navigation openings 61 m wide to beprovided

(b) two subsidiary navigation openings of 31 m(c) piers to be kept as narrow as possible, consistent

with their function(d) structures in the bed of the river in the navigation

openings to be at ruling depth to avoid restriction ofnavigation

IEE PROCEEDINGS, Vol. 134, Pt. A, No. 9, NOVEMBER 1987 753

(e) adequate overhead clearance to be provided in thenavigation openings for the passage of ships

(/) the reduction of the cross section of the waterwayby the permanent works not to exceed 25%, and by tem-porary works not more than 30%

(g) time required to close all gates not to exceed onehour

(h) the structure to be designed to withstand anextreme surge differential of 9.9 m and a reverse headdifferential of 6.1 m.In the early discussions with the navigation authorities, amain opening of 135 m was considered desirable inWoolwich Reach. This would have been flanked by twoopenings of 61 m [4]. When the closure of the SurreyDocks was announced, it was evident that the closure ofthe only remaining major enclosed dock system upriverof Woolwich Reach could not be long delayed. Therewere considerable advantages in dividing the 135 mcentral opening of the proposed structure into two 61mopenings. The gates and operating machinery of the fourmain openings could then be made identical and, in addi-tion, the failure to raise one of the four main gates wouldnot permit sufficient water under surge-tide conditions topass upriver and cause serious flooding.

Ships up to 20000 tonne GRT had passed through themain opening of Tower Bridge, 61 m in width, on manyoccasions with little difficulty, and, in the light of thisexperience, it was finally possible to persuade the naviga-tion authorities to accept the proposal.

For the 135 m main opening, a vast drop gate hadbeen proposed. The decision to provide four 61 m open-ings made the use of the rising sector gate possible. Theconcept of this type of gate is due to one of RPT's engi-neers, Charles Draper. This type of gate has been likenedto a huge beer can (25 m in diameter and 61 m high),with two thirds of the cylindrical shell cut away. If this

structure is placed horizontally in the navigation openingand the ends provided with passive pivots in their centreswith that part left of the cylinder barrel of the can on thebed of the river, ships would be able to pass freelythrough the opening. Closure of the structure understorm tide conditions is easily effected by rotating thestructure above its pivots, bringing the remaining strip ofthe cylindrical section up from the bed of the river until itcloses off the waterway (Figs. 3 and 4).

This type of gate has several advantages. It is robust indesign and reliable in operation, as closure is effected bysimple rotation of the gates. To provide for deflection ofthe gates under load, a hinged connection is fitted to oneend of the gate leaf, the other end being rigidly con-nected.

To meet the requirement that the area of the waterwaythrough the structure was not to be less than 75% of thenatural cross-section of the river, it was decided toprovide a further four openings, 31 m in width, threeclose to the north bank and one on the south bank,making ten openings in all. As these additional openingswere not for navigational purposes, but to allow free flowof the tide through the structure, normal falling radialgates could be provided in these spans.

Initially, it was proposed to rotate the rising sectorgates by means of winches and steel cables. To providean ample reserve of power, each set of machinery wasrequired to develop a pull of the order of 1000 tonne.This could have been done by means of multiple cables,but this arrangement would have created a maintenanceproblem, as the cables would have been subject to wearand corrosion. Other proposals for the operating machin-ery included drive-by sprockets, rotated by hydraulicmotors and engaging with toothed racks on the periph-ery of the gate ends, and ratchets reciprocated by hydrau-lic cylinders which engaged with a large toothed wheel

Fig. 3 An artist's impression showing the gates in the raised position

754 IEE PROCEEDINGS, Vol. 134, Pt. A, No. 9, NOVEMBER 1987

attached to the gate end. The final arrangement selectedwas a scaled-up version of the hydraulic mechanism usedto raise the blade of the conventional bulldozer.

A radius arm is caused to rotate by means of hydrauliccylinders acting on pins above and below the mainfulcrum. At the end of the radius arm, a link connects toa pin fixed to the gate end. The pull on the link generatedby the hydraulic cylinders causes the gate to rotate aboutits main pivot (Fig. 4).

radius arm

connectingsrods

link arm

shiftmechanism

hydraulicrams

sill

Fig. 4 Gate operating mechanism

To lock the gate rigidly in either the open or closedpositions and to enable the gate to be moved through asmall angle to permit flow to pass under the gate tosuppress any adverse effect downstream following barrierclosure, the shift and latch mechanism is provided foreach end of each rising sector gate. This consists of anarm provided with latches which clamp onto pins in theperiphery of the gate end. The other end of the arm isattached to aluminium bronze nuts on twin multi-startthreaded shafts which are rotated by hydraulic motors,the whole assembly being secured to the downriver endof the pier. Traversing the nuts along the shafts causesthe arm to move back or forth, so rotating the gate endthrough a small angle. This mechanism can also be usedto move the main mechanism through the 'dead-centre'position. Reversal of the main machinery then causes thegate end to rotate a further quarter turn, which bringsthe gate leaf clear of the water into the maintenance posi-tion. This mechanism is surprisingly powerful, developinga thrust of 350 tonne, the two sets giving a total thrust onone of the 61 m gates of 700 tonne. In an emergencythese mechanisms could be used to close the gate if neces-sary.

A further matter of considerable importance was thetype of bearing to be used to permit the rising sectorgates to rotate about the main pivots. These bearings hadto be capable of carrying a moving load of some 5000tonne and remaining in one position for long periods oftime without damage. In addition, they must operatesatisfactorily when flooded and be completely reliable.

Spherical roller, hydrostatic and solid lubricant bear-ings were considered. Of the three types, only the solidlubricant bearing met all the requirements, and so thistype was selected [5].

The main hydraulic cylinders which operate the mainmachinery for the 61 m gates are 1100 mm bore and3130 mm stroke. These are supplied with oil at up to17.24 N/mm2 by hydraulic power packs installed on eachpier. At maximum pressure, a thrust of 1500 tonne isexerted on the connecting rod linked to the radius arm.

In turn, a force of 900 tonne is applied to the pin on thegate arm.

Each pier with a 61 m gate on either side (piers 5, 6and 7) has three main power packs to supply oil at therequired pressure to the main cylinders. With normaloperation, a single power pack supplies oil to the set ofmachinery on one side of the pier. The third power packis installed as a standby.

Each of these main power packs has three electricallydriven pumps, a main pump driven by a 140 kW motorfitted with a constant-horse-power control unit. Themain pump has variable-stroke axial pistons fed by aboost pump, driven by a 22 kW motor which supplies oilat low pressure. The main pump delivers oil at pressuresup to 18.8 N/mm2 and flows up to 612 1/min. A smallpump, driven by a 2.5 kW motor, supplies oil for thecontrol valves. Similar, but smaller, power packs are pro-vided for the 31m gates.

Three sources of electrical power are provided for thesite. Each is independent of the others and is capable ofmeeting the peak demand for barrier operation. Two ofthe sources of supply are from the London ElectricityBoard, from separate connections to their networks onthe north and south sides of the river. One 11 kV connec-tion is made into a substation on the south bank, by anexclusive cable from a bulk supply station. Two connec-tions are also provided into the substation on the northbank from the local 11 kV ring main distribution system.The third source of supply is from diesel engine drivengenerators installed in a generating station on the southbank. Three 2000 h.p. turbocharged 16 cylinder dieselengines drive 1.5 MW alternators generating at 11 kV.For normal operation, two sets each feed half the barrierload (peak demand 2.7 MW) with the third set onstandby. In an emergency, with suitable load sheddingone set is able to provide sufficient power to close thebarrier.

Power is distributed at 11 kV from the shore sub-stations, through the duplicated service and accesstunnels, to each pier and the north and south abutments.Each pier or abutment has two transformer stations con-nected to medium-voltage switchrooms on the upriverand downriver ends of the pier or abutment.

Connections from the north and south substations tothe transformer stations on the piers are 'handed' so thatpower is available to each pier, even in the event ofdamage to one of the access tunnels or failure of two ofthe power sources (Fig. 5). Operation of the Barrier isnormally from the main control room on the south bank,but can be carried out from the local control rooms onthe piers [4].

3 Construction

Three main contracts and some twenty other contractswere let for the construction of the barrier. The maincontracts covered the civil engineering work in the con-struction of the north and south abutments, the ninepiers in the river and the sills in the river bed; the struc-tural work in the fabrication and erection of the gates;and the manufacture and installation of the operatingmachinery. Other contracts were let for the clearence ofthe site; construction of access roads and offices; dredg-ing of the diversion channel in the northern half of theriver; construction of the control building, workshop andgenerator house; installation of mechanical and electricalservices; generating plant; and the supply of switchgearand transformers.

1EE PROCEEDINGS, Vol. 134, Pt. A, No. 9, NOVEMBER 1987 755

The decision was taken to start work on site on thesouthern half of the structure first. A diversion channelwas dredged in the northern half of the river to enable

upriver tunnel

pressure, made excavation and construction inside thecofferdam very difficult. For the construction of theremaining northern piers following the switch of naviga-

1 °9 CD

pier pier pier pier pier pier

5 4 3

tunnelsou th shoresub stat ion

Fig. 5 11 kV distribution system

shipping to pass the site, as the work in the southern halfencroached on the existing shipping channel. Thisentailed the removal of about 106 m3 of material and thework was completed by the end of January 1975.

The main civil engineering contract, £38 x 106 invalue, was awarded in the summer of 1974 to a consor-tium of Costain Ltd., Tarmac Ltd. and Hollandsch BetonMaatschappij BV. The work on site on the south abut-ment and the four southern piers started early in 1975.Work was also started on the north bank, where a largebuilding dock for the construction of concrete sills, forthe southern navigation openings, the north abutmentand the two most northerly piers, was constructed in theshallow water close to the north bank. It was intendedthat once the southern piers were sufficiently advanced,shipping would be diverted through the two southern61 m openings, permitting work to go ahead on theremaining northern piers.

Several methods were considered for the constructionof the piers, but the one finally selected by the CTH con-sortium was the traditional steel-sheet pile cofferdam.The design of the structure required the founding of thepiers on the chalk some 25 m below mean tide level,except in the case of the three most northern piers whichwere founded on the Thanet Sand. The cofferdams for themain piers flanking the 61 m openings were very large,75 m x 17 m internally. Larssen number 6 piles wereused, some 30 m in length, requiring four strutting framesto resist the water pressure. Excavation was carried outunder water, until the foundation level on the hard chalkwas reached. A 5 m thick base slab of concrete was thenplaced underwater, to stabilise the cofferdam, and also toseal the floor against the inflow of water from fissures inthe chalk. Once this was in place, dewatering of the cof-ferdam could take place. Pressure relief wells were formedin the concrete slab by casting a number of 865 mmdiameter steel tubes in the slab and allowing water fromthe underlying chalk to escape into the cofferdam, and toflow to the perimeter drainage system.

The heavy strutting necessary to provide this design ofcofferdam with the strength required to resist the water

756

north shoresub station

tion through the southern half of the barrier, a newdesign of cofferdam was developed. This made use of aheavy German pile, the Peine pile which took the form ofa heavy steel joist 900 mm deep. These were driven withthe web of the pile at right angles to the line of the cof-ferdam wall, the flanges of the piles forming the inner andouter faces of the cofferdam wall. Special linking barsconnected the flanges, together providing a watertightface. These piles were sufficiently strong to require onlyone strutting frame which was not put in position untilafter most of the excavation had been completed. Eachpair of piles weighed some 20 tonne and were driven bythe heaviest diesel piling hammer available, the DelmagD62. The final stage of driving these piles into the hardchalk required a thousand blows per metre of penetration[6] (Fig. 6).

Once the piers in the southern half of the river weresufficiently advanced for the cofferdams to be removed,the float in of the sills, which had been constructed in thenorthern building dock and had been stored temporarilyin the Royal Docks, could take place. This was a difficultoperation, as the 61 m sills weighed some 10000 tonneand the clearance between each end of the sill and thepier was only 50 mm. In addition, tidal range was about4 m even on neap tides. The operation was carried outsuccessfully in every case as result of careful planning andconsiderable research work.

By June 1980, the civil engineering work on thesouthern piers was sufficiently advanced for access by thestructural and mechanical engineering contractor. Asalready mentioned, the original intention was to letseparate contracts for the construction and installation ofthe machinery and the fabrication of the gates. However,the most favourable tenders for both contracts werereceived from the Davy Cleveland Barrier Consortium, ajoint venture by the Davy Loewy Company of Sheffield(now Davy McKee) and the Cleveland Bridge Companyof Darlington. A combined contract was therefore let, fora sum of £28 x 106.

Cleveland Bridge was responsible for the fabricationand erection of the gates and the erection of the machin-

IEE PROCEEDINGS, Vol. 134, Pt. A, No. 9, NOVEMBER 1987

ery on site. Davy Loewy manufactured and assembledthe machinery, largely at Sheffield, but subcontracts wereplaced with firms all over the country.

Fig. 6 Aerial view of work in progress

To make the maximum use of their workshops at Dar-lington, Cleveland Bridge divided each of the main gatesinto five strips running from end to end, two edge stripsand three internal strips. Each strip was divided into fourparts, giving sections of gate 15 m long, 5 m high and2 m wide. The weight of these sections was some 50tonne, and they were transported by road from Dar-lington to the erection site on the north bank of the RiverTees at Port Clarence. Here the remaining plates to com-plete the internal diaphragms and the outer skin, and theinternal struts were welded in place. When complete andwhen the site at Woolwich was ready to receive them, thegates, gate ends, rocking beams etc. were transported bysea on large barges to the site.

Many specialist suppliers and subcontractors contrib-uted to the machinery contract. The hydraulic cylinderswere manufactured by Henry Berry of Leeds, and thehydraulic power packs by Vickers of Eastleigh. Fabrica-tion of the radius arm bearing brackets was by Weldall ofStourbridge. Assembly and testing of units such as theshift and latch mechanism took place at Sheffield. Themain trunnion shafts, about which the gates rotate, weremanufactured under a separate contract placed with theBritish Steel Corporation, and the trunnion shaft supportstructures which were built into the piers, to which thetrunnion shafts which were bolted were manufactured byVoest Alpine AG. A trial assembly of these two units wasarranged at Sheffield to ensure that there were no prob-lems later on site at Woolwich when it was necessary tobolt these two units together. The main bearings aboutwhich the gates rotate were manufactured by Merrimanof Boston, USA, and were fitted to the gate arms at PortClarence, together with the trunnion shafts, before thearms were transported to the Thames.

Work on site took place over three years from April1980, when access to the first pier was possible. Access forgate erection was possible from July 1980 and loading

out and transport of the gates and gate ends then wentahead. The proposed plan for gate erection followed alogical plan. Work was programmed to start with theerection of the 31m falling radial gate on the south sideof the river, followed by the first 31m rising sector gate.The most southern of the 61m gates would then follow.

Gate ends complete with bearings and trunnion shaftswere to be placed first by floating cranes. The 1200 tonnegate ends were lowered onto underwater hydraulic jacksset in recesses in the piers. Adjustment of the jacks wouldallow the gate end to be moved into position with therequired degree of precision, an accuracy of location of1^ mm was required. The gate ends were to be placedwith the bolting face for the gate leaf uppermost, so that,subsequently, the gate could be floated in, lowered ontothe bolting face and the bolts put in place.

Erection of the first falling radial gate and the first31m rising sector gate went ahead without problems. Inboth cases, the loads were well within the capacity of asingle floating crane. The first of the gate ends for the61 m rising sector gate was also placed successfully. Thisoperation entailed the use of two 800 tonne capacityfloating cranes, working in tandem to handle the 1200tonne load. The gate end was in the horizontal positionon the barge used to bring it from Port Clarence, andhad to be raised and swung up into the vertical positionby the use of a special spreader beam and auxiliarytackles to enable it to be offered up to the pier.

Difficulty was experienced with the placing of thesecond gate arm on pier 7, for the most southerly of the61 m spans. It was not possible to co-ordinate the lower-ing operation by the two cranes sufficiently, to stop thegate end rotating on the jacks as it was lowered. This, inturn, made it impossible to attach the links used to securethe top of gate end to the pier. After several attempts, itwas decided to abandon the operation and to have afresh look at the method used for placing the gate ends.The underwater jacks had been set radially and the con-clusion was reached that this was the root of theproblem. It was decided to set the jacks with their axesvertical, and to weld lugs onto the gate ends which wouldrest on the jack heads and transfer the weight of the gateend to the jacks. The two 600 tonne jacks originally pro-vided were replaced with four coupled 250 tonne jacks,each with spherical head bearings, to ease handling andto increase reliability. The gate end on the south side ofpier 7, which had not been possible to place in position,was 'parked' temporarily on packing on the pier and tiedback to the pier face. Availability of the heavy lift cranesand the progress of the civil engineering contract led to adelay of some twelve months before the new method ofplacing of the gate ends could be tried out. This provedsuccessful and, by reprogramming the sequence of gateinstallation, it was possible to achieve the originalplanned operational date [5].

The first gate leaf was placed in position in the spanbetween piers 6 and 7 on 14th December 1981. The oper-ation went much as planned, but not without some lastminute misgivings, this time on account of the weather. Adeep depression moved in from the West during 13thDecember, bringing gale force winds and a blizzard. Itlooked very much as if the operation, scheduled for the14th December would have to be cancelled. Last minuteinformation from the Meteorological Office held out aray of hope however. The centre of the depression wasexpected to pass over the site during the morning of the14th and there should be a period of still air and sun-shine. The forecast proved correct, the wind dropped and

IEE PROCEEDINGS, Vol. 134, Pt. A, No. 9, NOVEMBER 1987 757

the sky cleared. By 09.00 h the sun was out and theplacing of the gate leaf could go ahead. By midday, thefirst gate leaf was in position on its two gate ends. Oncethe technique had been mastered, the placing of theremaining 61 m gate ends and gate leaves went ahead toschedule.

Installation of the operating machinery went ahead atthe same time as gate installation. Only limited accesswas possible through the service tunnels in the gate sills,so that most of the units and materials were transportedby water. The large hydraulic cylinders were placed onthe piers by floating crane and were then moved intoposition with skates. Testing and commissioning wascarried out methodically in stages, and few problemsarose.

By the end of October 1982 it was possible to close allthe gates together at low tide for the first time, and aweek later a further closure was carried out on the risingtide, which developed a differential head across the struc-ture and put the various units under load (Fig. 7).

4 Other contracts

Besides the major contracts for the civil engineering workand the machinery and gates, some twenty other con-tracts were let for various sections of the work. The con-tracts for the trunnion shafts, the main bearings and thetrunnion-shaft support structures have already been men-tioned.

Other contracts were let for the buildings on the southbank; the diesel generators; high-voltage switchgear andtransformers; control panels and equipment; mechanicaland electrical services; local control panels; low-voltageswitchgear; lifts; navigation lights and ancillary works onthe south bank.

5 Project management aspects

The management of the project passed through threestages, from the start of the Greater London Councilsinvestigation in 1968 to its completion in May 1984. Thefirst phase was the organisation set up to deal with thefeasibility of the project and the selection of the preferredsolution; the second phase, the development of the detaildesign and the drafting and letting of the contracts forthe construction; and, the third phase, the constructionand commissioning of the structure.

For the investigation stage, the GLC set up a smallbarrier project team. Policy was laid down by a policycommittee which co-ordinated the views on policy for theproject from the Ministries of Agriculture, Fisheries andFood; Housing and Local Government, and other Gov-ernment Departments with those of the GLC. A steeringcommittee then monitored the work being initiated bythe project team. A number of working parties were setup to enable the project team to draw on the best exper-tise available in the country, in the fields of oceanogra-phy, navigation, construction etc. Research organisationswere called on to assist in areas where some degree ofuncertainty existed. A joint organisation of two wellknown firms of consulting engineers developed outlineschemes for a number of proposals. Fig. 8 illustrates dia-grammatically this organisation.

Agreement on the preferred scheme resulted in theGLC obtaining powers from Parliament to construct thebarrier and the transfer of responsibility for grant aid forthe project to the Ministry of Agriculture, Fisheries andFood. Executive authority for the work was delegated bythe GLC to their Public Services Committee who wereresponsible for approval of estimates, authority to spend,letting of contracts, and payments for execution of work.

Preparation of estimates, reports, contract documents,supervision of work and certification of payments to con-

Fig. 7 Completed barrier closed against the rising tide

758 IEE PROCEEDINGS, Vol. 134, Pt. A, No. 9, NOVEMBER 1987

tractors and consulting engineers were the responsibilityof an enlarged GLC project team. Messrs. Rendel,Palmer and Tritton were appointed the Council's con-sulting engineers to carry out the detail design of the

Ministry ofAgriculture,

FisheriesandFood

Ministry ofHousing and

LocalGovernment

PolicyCommittee

otherGovernmentDepartments

SteeringCommittee

Research Organisations

HRSIGS105BHRA

Working Parties

Navigation ThamesBarrierprojectteam

GLC Departments

Pollution and SiltationAmenityOcean and Met.Civil Engineering _

Fig. 8 Project management organisation structure

barrier, drafting contract documents for the barrier andsupervision of work in liason with the project team. TheMinistry of Agriculture, Fisheries and Food, as the Gov-ernment Ministry responsible for the project, required, asa condition of payment of grant, detailed consultation onevery aspect of the work, from design, contract condi-tions, specifications, selection of contractors, letting ofcontracts, variations, payments and settlement of claims.

Rendel, Palmer and Tritton appointed a projectmanager, who, with his design team, worked very closelywith the GLC team, virtually as one organisation.Monthly progress meetings were held where design prob-lems were discussed, progress monitored and action initi-ated to deal with difficulties. A comprehensive but verycomplicated critical path network was set up, but as timewent on this became very time-consuming and a simpli-fied version was used.

The main barrier contracts were let in the summer of1974, and the work moved on to the construction stage.Later contracts were still in the design stage, however, sothat both aspects of the work were in progress at thesame time.

The project management organisation had therefore todeal with both types of activity concurrently. No markedchange in the structure was necessary, but additionalcomponents were added (Fig. 9). A monthly site meetingbetween the consulting engineers project manager, sitestaff and the GLC's project manager took place, dealingwith the progress on all contracts. Separate monthly sitemeetings also took place for the main civil, mechanicaland structural contracts, with representatives of the con-tractors involved.

A major problem was the delay on the main civilengineering contract due to low productivity. This led tostorage problems for items of the gate and machinerycontracts, until the site was ready to accept them. Initi-ation of the manufacture of some items was delayed, butdelay costs on the civil engineering contract would havebeen so high that no risks were taken in this respect.

Ministry ofAgricultureFisheriesand Food

GLCPublic

ServicesCommittee

GLCproject

team

bankraisingteams

GLCDepartments

consultingengineers

projectmanager

otherbarrier

contracts

researchorganisations

projectmanager

civi l

projectmanager

mechanicaland structural

contractengineering

contract

Fig. 9 Project management structure with additional components

6 Costs

The total cost of the barrier escalated from £110 x 106,when all the main contracts based on 1973 prices were letin 1974, to £440 x 106, when the work was completed in1984. 70% of this increase was due to inflation, as all themain contracts were index-linked using either the Baxteror BEMA indices. The remainder of the increase was duelargely to low productivity, particularly on the civilengineering contract. The total quantity of materialshowed little increase over the original quantities, butlabour costs increased, as did plant costs due to theincreased duration of the work.

7 Conclusions

The mangement of the project was complicated by thenumber of authorities and organisations involved andalso by the transition from feasibility stage throughdesign to construction. As a result of clear definition ofareas of responsibility and an understanding by allinvolved of the 'rules of the game', the organisationworked surprisingly smoothly. It was impossible for anyone person to be in the position of the all powerfulsupremo beloved by management theoreticians, because,

IEE PROCEEDINGS, Vol. 134, Pt. A, No. 9, NOVEMBER 1987 759

by law, contracts could only be signed, costs incurred andbills paid with the approval of the Greater LondonCouncil itself, which, in turn, because of the Govern-ment's grant aid conditions, was dependent on approvalby MAFF and even the Treasury on major items ofexpenditure. In practice, decisions were taken at variouslevels according to the nature of the decision. Decisionon technical matters was the responsibility of the consult-ing engineers with the approval of the GLC and MAFFengineers. Contractual matters involving increased costsmight have to be referred upwards with the final wordcoming from the Treasury. On no occasion was there animpasse or even undue delay, due very largely to thegood sense and co-operation of all concerned.

8 References

1 BO WEN, A.J.: 'The tidal regime of the River Thames; long termtrends and their possible causes', Phil. Trans. R. Soc. London A, 272,pp. 187-199

2 BONDI, H.: 'London Barrier'. Report to the Minister of Housingand Local Government, 1967.

3 Report of the Departmental Committee on coastal flooding, Cmd9165 (HMSO, 1954)

4 Thames Barrier Design' (Institution of Civil Engineers, London,1978).

5 'Thames Barrier'. Institution of Mechanical Engineers, Papers pre-sented at a meeting at Institution Headquarters, 8 June 1983

6 GRICE, J.R., and HEPPLEWHITE, E.A.: 'Design and constructionof the Thames Barrier cofferdams', Proc. Inst. Civ. Eng., 1983, Part 1,74, pp. 191-224

760 IEE PROCEEDINGS, Vol. 134, Pt. A, No. 9, NOVEMBER 1987