Revised manuscript for World Steel Bridge Symposium Nov.

2009 in San Antonio References and small corrections have been added. Some of the references will be in place later. The full text of the references can be found in the list of literature in The Network Arch on my home pa e! http!""pchome. rm.hia.no"#ptveit" $nclined characters are used when $ was supposed to point at a slide.

1 Networ arc!" #angama!u. %y name is &er Tveit. $ come from Norway. The title of this lecture is! 'enesis and development of the network arch. The network arch is an extremely li ht arch brid e that $ came to think of when $ was workin on my master(s thesis )* years a o. $ am very thankful for this opportunity to tell you about it. $ have a lot to tell+ still there will be room for your ,uestions at the end of the lecture. Network arches are arch brid es where some han ers cross other han ers at least twice. Touchin on the development of network arches+ $ want to tell you how network arches function+ about their optimal form and how they can be erected. Some examples of built network arches will be presented. %etric units will be used. $ am sorry if that is a problem.

2 Nielsen Bridge. The forerunner of the network arch was the Nielsen brid es in Sweden. -Nielsen ./0/1 Their chords were so stocky+ and the ratio of live load to dead load was so small+ that they did not need crossin han ers to achieve small bendin moments in the chords. $n a patent application from
./02 Nielsen showed han ers that crossed other han ers once. $f built these brid es would not have been network arches. $n the brid es that Nielsen built+ han ers did relax due to loads on part of the span. Please note that there are no railings between the hangers and the traffic.

$ Beam Analogy. $ncreased loads and stron er materials make it advanta eous to use han ers that cross each other at least twice. The network arch can be seen as a simply supported beam. The chords are the compressive and the tensile 3ones. The han ers are a very li ht web. % &artial continuation 1. %ost of the shear force is taken by the vertical component of the axial force in the arch. The han ers take some of the variation in the shear force. $ncreased rise of the arch could ive much smaller axial forces in the chords+ but it would detract from the ood looks of the brid e.

' &artial continuation 2. The arch should normally be part of a circle. The nodes can be placed e,uidistantly alon the arch or alon the tie. The arch is well supported and can efficiently utili3e hi h stren th steel. That also applies to all the other steel members. As lon as the han ers remain in tension+ the bendin moments in the chords are very small. 4hen han ers relax+ the bendin moments in the chords increase.

( )oncrete ties wit! diagram of t!ic nesses. $f the arches are less than .) m apart+ it is best that the tie is a concrete slab with small ed e beams. Normally the biggest bending moment in the tie is in the middle of the slab. The lon itudinal bendin moment in the tie is normally smaller. The axial force in the tie is best taken by prestressing cables in the edge beams. 5or everyday loads the prestressin cables ive a lon itudinal prestress in the concrete tie. This makes the tie more durable. $t mi ht be a bit like swearin in church to defend the concrete tie of network arches in this conference on steel brid es. $ humbly apolo i3e+ but the concrete tie is a cost6 efficient solution when the arches are less than .7 to .) m apart. Then we do not need steel beams in the tie.

8on itudinal steel beams in the tie would lead to extra reinforcement to reduce the crack width in concrete that rests on the elon atin lon itudinal steel beams. &restressin cables combined with lon itudinal steel beams in the tie would introduce unfavourable compressive stresses in the steel beams due to creep and shrinka e in the concrete slab. 8on itudinal steel beams in the tie reduce the bendin moments in the arches and increase their bucklin stren th+ but these effects would be sli ht and unimportant. 8on itudinal steel beams in the tie can be very useful in many methods of erection.

* +ransverse tension mem,ers. 4hen hi h stren th concrete is used to obtain better concrete durability+ we can make the concrete slabs thinner and hope that the deflections will turn out as computed. $f we et too bi deflections+ they can be counteracted by transverse tension members under the tie. - Wedges under tension mem,ers. Over the years, the size of the deflections can be controlled by putting wedges between the slab and transverse tension members. $f the bendin moments can be taken by the reinforcement in the slab+ rupture of transverse tensile members will not endan er the brid e.
$f the transverse prestressin consists of replaceable steel rods like in 'ermany+ their stress can be altered to ad9ust the deflection of the slab. Then the span of the slab can be over .) m.

9 Sc!ulen,urg Bridge" cross section. :ven if we have transversal steel beams in the tie+ the lon itudinal steel beams in the tie can be avoided. ;ow this can be done is shown in this slide of the Schulenbur <rid e. $t has a span of /7 m. The brid e was the sub9ect of 4olfram <eyer=s master(s thesis at the >resden technical university. ;is adviser was &rofessor >r. 5rank Schanack+ who is present at this conference. ;e is probably the man in the world who knows most about network arches. The transversal steel beams in the tie are deeper than the lon itudinal concrete beams.

10 Sc!ulen,urg Bridge" picture. This is a picture of the Schulenbur <rid e. The deep beams do not make the brid e look clumsy+ but it would have been more ele ant if the tie had been half a metre deep instead of ..?? m deep. 11 .e!marn Sound Bridge. -Stein and 4ild ./2)1 $f there are transverse beams+ the han ers would normally be fastened to the ends of these beams. Then the han ers would have many different diameters. $n that case there are no stron reasons to avoid a constant slope of the han ers. This arran ement has been used in the 5ehmarn Sound network arch in 'ermany and in many @apanese network arches. -Nakai .//)1
12 Stein /er. -Tveit 077?1 The optimal han er arran ement depends on many thin s! ratio of live load to dead load+ ratio of concentrated to evenly distributed load and len th of concentrated load and the rise of the arch. The arch is normally part of a circle+ maybe with a sli ht deviation near the ends of the arches. $t is advanta eous to have little variation in the han er diameter. :specially when the han ers are fabricated locally as was the case for the first two Norwe ian network arches. They had a constant chan e of slope between ad9oinin han ers. The maximum han er force was less than .7 A bi er than the avera e han er force.

1$ Stein /er influence lines. This slide shows the influence lines of the Steink9er network arch. -Tveit ./221 ou can see that the axial forces in the chords must have small variations. The same applies to the maximum bending moments in the tie. The slope of the han ers is stated in the top ri ht corner of the slide. 1% 200 m !anger arrangement. The han er arran ements in this slide were found by trial and error in ./?/. $f there are no transverse beams in the tie+ then it is usually best to use a constant distance between the nodal points in the arch. The han er arran ement to the ri ht was used in Teich and 4endelin(s master(s thesis in 077..
1' 0 vi Sound 2001. This slide shows the network arch in Teich and 4endelin(s master(s thesis. -Teich and 4endelin 077.1 :B loads and codes were used. The tie is a concrete slab. The arches are made of universal columns with a yield stren th of */7 N"mm0. The universal columns are supposed to come pre6bent from the steel works. Two ways of fastenin the windbracin are shown. 1( Steel weig!ts in various arc! ,ridges. $n this slide the steel wei ht in Teich and 4endelin(s network arch is compared to steel wei hts in some 'erman tied arch brid es with vertical han ers. N means that there is no windbracin . ! means that the arches slope towards each other. The years when the brid es were built are indicated.

The network arch has no steel beams in the tieC still it needs about the same amount of reinforcement as the other brid es. The slide supports my claim that network arch brid es can save over D of the structural steel needed in other steel brid es.
1* 1er2og. This slide compares network arches to other steel brid es that have been built. The other brid es are from before ./?E. $ believe that today they would have used more steel. Fptimal two6lane network arches would use sli htly less steel than indicated by this dia ram.
-;er3o ./?)1

1- &oints of importance. Gou all know that steel wei ht is not the only tin that matters. This slide indicates other thin s of importance. Network arches are slimmer. That looks ood. They have a thinner lower chord. That makes ramps shorter. $t also makes it easier to desi n the roads leadin up to the brid e. Fptimal network arches have shorter welds and simpler details. Fther tied arch brid es have much more surface that needs corrosion protection. Fther concrete parts need much more maintenance than concrete slabs with a sli ht prestress. :rection is often more expensive with two to four times more steel to erect.

%aybe the structural steel for Teich und 4endelin(s network arch would not cost more per tonne than the steel in the 'erman tied arch brid es. The steel for the arch is #H7 A the wei ht.
19. Steel (0 to 200 m. This slide ives a rou h estimate of steel wei hts in two6lane network arches. The spans are between 27 and 077 m. 8on er spans mi ht not be ,uite so competitive+ because the crossin han ers reduce the stresses due to live loads+ but not the stresses due to permanent loads. The dots indicate the steel wei hts from Teich and 4endelin(s Ikvik Sound <rid e. The dia ram can be used to compare steel wei hts of proposed brid e alternatives.
20 B3S railway ,ridge. The network arch is very well suited for railway brid es. <runn and Schanack(s master=s thesis was the desi n of this two track railway brid e in 077E. -<runn and Schanack 077E1. $t has a span of a .77 m. 21 +ie. This slide shows the reinforcement in the tie in <runn and Schanack(s railway brid e. $t uses very little steel. 22 Steel weig!ts. This slide shows the steel wei ht pr m track in many railway brid es. Gou can see that <runn and Schanack(s brid e uses about J of the steel needed for the other railway brid es. 2$ Rail (0 to 200 m. ;ere Schanack ives a rou h estimate of steel wei ht in two track railway brid es. The spans are between 27 and .)7 m. The dot indicates the steel wei ht found in <runn and Schanack(s master=s thesis in 077E.

The dia ram can be used for comparin the steel wei hts of two track railway brid e to other alternatives. The 'erman railway(s advisory board for brid e desi n favours network arches over arch brid es with vertical han ers. -Schlaich et al+ 077H1. 5our network arch railway brid es have been built in 'ermany. 2% Sc!anac 1. 5rank Schanack has made two very interestin dia rams that show the advanta es of network arches. -<runn and Schanack 077/1 $n this dia ram he compares a network arch to a tied arch railway brid e with vertical han ers. The span is a hundred meters. $n the upper third of the dia ram+ bendin moments due to dead loads are shown.
<ecause the arch is not a parabola+ but part of a circle+ the bendin moments in the arch with vertical han ers are .7 to .) times bi er. The middle of the dia ram shows the bendin moments due to maximum live loads. The maximum bendin moment in the tie is .. times smaller in the network arch. $n the arch of the network arch+ the bi est bendin moment is around .7 times smaller. The lower third of the dia ram has maximum live load on half of the spans. The maximum bendin moment in the tie is 07 times smaller in the network arch. $n the arch of the network arch+ the bi est bendin moment is around .E times smaller.

2' Sc!anac 2. $n this dia ram Schanack starts at the top left with a tied arch with vertical han ers and the maximum live load on half the brid e as shown in the previous dia ram. Then the an le between the arch and the han ers is radually reduced. At the bottom left of the slide the maximum bendin moment is only 2A of the startin value.
2( .ran 4s !anger arrangement. $n 077E 5rank Schanack su ested a constant an le between the arch and the han ers+ except near the ends of the arches. This han er arran ement has many advanta es.
2* Scaffolding for Bolstadstraumen. Now to the erection of network arches. Network arches have been built or are bein built in 'ermany -'raKe 077?1+ &oland -Loltowski 077)1+ M3ech Republic+ -Sasek 077)+ 07721+ Slovakia+ Norway+ Spain+ The Bnited States+ -Steere 077H1+ -4ollmann 077H1+ &eru+ Ar entina+ New Lealand -Mhan 077H1+ @apan+ -Nakai .//)1 Taiwan and maybe more places. :rection is done in many ways. The two Norwe ian network arches were erected on a timber structure restin on piles in the river bed. The slide shows the scaffoldin for the <olstadstraumen network arch in Norway. 4hen the scaffold was finished+ the concrete tie could be cast. Then the arches were erected and the han ers were put in. Then the han ers were ti htened until the arch carried the tie and the timber structure could be removed.

2- +emporary lower c!ord. This temporary tie can be used for the erection of network arches. Mombined with arches and han ers it makes a stiff steel skeleton that can be moved. $t can carry the castin of the tie. 5irst the concrete at the ends of the span must be cast. Then the ed e beams. They mi ht have to be cast from both ends to avoid excessive relaxation of han ers. 4hen the ed e beams have been cast+ they take the bendin in the tie while the concrete lane is cast. N$ could have said that this is the steel tie of a
network arch. $t can be removed when the concrete can stand by itself. Then corrosion protection for the steel tie can be omitted.O

29 5ifting t!e steel s eleton. The steel skeleton of the network arch from Teich and 4endelin(s master(s thesis can be lifted by Norway(s bi est floatin crane. The span was .E) m.
The liftin capacity of the crane is 277 t. The steel skeleton wei hs around 0E7 t+ but it is practical to put in some of the wooden form and the reinforcement before the skeleton is lifted to the pillars. $0 Removing t!e temporary c!ord. This wa on can be used for removin the temporary tie and the formwork. Part of the deck in the wagon has been a horizontal part of the form for casting the concrete tie . The wa on rolls alon the ed e of the finished tie. $f there is transverse prestressin of the tie+ the stressin can be done from the wa on. The same temporary lower chord can be used for brid es with various spans and widths.

$1 S ew Bridge. This slide shows the first sta e in the erection of a network arch across a canal. The span is a hundred metres. $n order to reduce the necessary thickness of the concrete tie+ the brid e has three arches. The structural steel supplemented by a temporary lower chord is erected at the side6spans on one side of the canal. $f the shape of the steel skeleton is ri ht+ then no ad9ustment of the han ers will be necessary later. 4hile the steel beams on top of the pontoon are tied to the abutment+ the steel skeleton is rolled to the middle of the pontoon. Then the pontoon is pulled across the canal. 5inally the steel skeleton is rolled onto the abutment on both sides of the canal. Then the tie is cast.
$2 #angama!u. This slide tells about the erection of the %an amahu <rid e in New Lealand. -Mhan 077H1 The steel tie was erected on two temporary supports. Two halves of the arch with windbracin were erected by the two red mobile cranes. The yellow crane in the picture carried the two men that 9oined the arches at the top. Afterwards the han ers were put in. Then thin concrete plates were placed on top of the steel tie and the concrete slab was cast. $f we define the slenderness of an arch brid e as the span divided by the combined hei ht of the chords+ this is the world(s most slender tied arch brid e. $t ot a old award of excellence in New Lealand this year.

$$ Bec!yne p!oto. This slide shows a network arch that was built in the M3ech Republic in 077*. -Sasek 077)+ 07721 $t has a span of *. m. $t replaced an old brid e with too little room for the .77 year flood.

The existin roads on both sides of the brid e needed only small alterations. The M3ech en ineer 8adislav Sasek heard about the network arch. Then he consulted my home pa e and desi ned this brid e. $% Bec!yne )ross6section. This is a cross6section of the <echyne <rid e. &lease note how the pipes are hidden under the footpath. $' Bec!yne lifting. Sasek su ested this method of erection for the steel skeleton of the network arch in <echyne. The steel skeleton wei hs about E7 t. 07 t of the load is formwork and reinforcement. 8adislav Sasek is now busy buildin a 077 m network arch that won a desi n competition in &ra ue. $( &rovidence. 7eneral drawing. This network arch in &rovidence in the BSA was finished in 077?. -Steere 077H1 The span is .00 m. $t has three arches. There are five lanes in each direction. The slope of the han ers is 27P. There is only one han er at the end of each transverse beam. $* &rovidence" Steel s eleton. This slide shows the 0)77 t steel skeleton of the &rovidence network arch before it was floated 07 km to the site. $- Blenner!assett Bridge. This slide shows the mail span of the <lennerhassett <rid e across the Fhio River. -4ollmann 077H1 $n was opened in 077?. The span is 02H m. There is E0.2 m between the planes of the arches. Mompared to a brid e with vertical han ers+ the crossin han ers reduce the live load deflections by a factor of ...

$9 Blenner!assett erection. The method of erection of the <lennerhassett <rid e was influenced by the need to have room for passin ships. The slide shows erection when one of the last beams in the tie was about to be put in place. The brid e was chosen because it was less costly than a previous desi n. This year it ot a national award for lon span brid es.

%0 )onclusions after '% years. &art 1. After )* years with network arches+ $ have arrived at some conclusions that $ will present in the next three slides. A well desi ned network arch is likely to remain the world(s most slender tied arch brid e. The slim chords are pleasin to the eye and do not hide the landscape or cityscape behind them.
The slim tie is an advanta e when the brid e is lifted up to let traffic pass under it. $t also makes ramps leadin to the brid e shorter. The small vertical reactions ive savin s in the substructure. $f the distance between the arches is less than .)607 metres+ a concrete tie would often be a ood solution. Moncrete ties should have small ed e beams with room for the lon itudinal prestressin cables. $f the distance between the arches is more than .7 m+ transversal prestressin should be considered. Network arches have small lon itudinal bendin moments in the chords.

%1 )onclusions after '% years. &art 2. Network arches are e,ually well suited for road and rail brid es. All their members efficiently carry forces that can not be avoided in any simply supported beam. Tie and han ers ive the arch ood support and hi h bucklin stren th in the plane of the arch. Tension is predominant in tie and han ers. ;6beams with hori3ontal webs NBniversal columns or American wide flan e beamsO have a favourable distribution of stiffness and ive simple details. All han ers can have nearly the same cross6section. :fficient methods of erection are available. :rection can be done usin a temporary lower chord which+ combined with the structural steel+ has enou h stren th and stiffness to carry the castin of the concrete tie.
%2 )onclusions after '% years. &art $. Network arches are not sensitive to uneven settlements in the foundations. ;i h stren th and low wei ht ive the network arch ood resistance to earth,uakes. %ost concrete parts need more maintenance than a concrete slab with a sli ht prestress. Network arches have small surfaces and need little corrosion protection. ;i h stren th steels are well utilised. $f thin s o well+ the network arch can save up to *7 A of the cost and ?7 A of the structural steel. NTed Loli saved more than )7A of the cost by usin two network arches instead of a cable stayed brid e footbrid e in San @ose. The spans were the same.O A hi h percenta e of the cost will be labour. Monservatism and lack of time are important obstacles to the buildin of network arches. Thank you all for listenin . $ look forward to your ,uestions now and durin the whole con ress. $f you want me to+ $ can lead you to the slides and the manuscript for this lecture. 6 Thank you very much.
This manuscript+ the slides and my contribution to the 077/ 4orld Steel <rid e Symposium can be found at