The eventful history of the Carola Bridge in Dresden

First of two parts: Dresden Elbe bridges, supporting structure and construction work

The collapse of the Carola Bridge on September 11, 2024 is certainly the most significant case of corrosion-related material failure in recent times and also serves as a prominent example of the country's ailing infrastructure. The partial collapse of the prestressed concrete bridge in Dresden occurred without prior notice. Originally, the Dresden Carola Bridge was considered an icon of engineering of its time. It was a very aesthetic and slender structure. Its design and construction would still be a challenge today. The existing documents testify to a high quality of design and very careful construction. The first part of the article presents the history, design and construction of the bridge. The second part in the June issue deals with the collapse, the attempt at reconstruction and the search for the cause of the collapse.

The Elbe bridges in Dresden city center

It is still not known exactly when the oldest bridge over the Elbe in Dresden was built. The first mention is dated to the period between 1228 and 1234, although the relevant source specifically mentions a repair but not a (new) construction. This first stone bridge was around 560 m long and 8.5 m wide. The interior of the 24 arches consisted of quarry stones embedded in lime mortar, which were clad on the outside with sandstone ashlars. For many centuries, the structure was considered one of the most monumental and longest stone bridges of the Middle Ages [1].

Before the Second World War, four arched bridges connected the Neustadt and Altstadt sides of Dresden for road traffic.

In the direction of flow of the Elbe (from east to west), these were the König-Albert-Brücke (1877, today Albertbrücke), the Königin Carola-Brücke (1895), the Augustusbrücke and the Marienbrücke (road bridge from 1852, parallel to which a railroad bridge has existed since 1901), see for example various articles in [2] and [3]. All these bridges were partially blown up by German soldiers in the last days of the Second World War in May 1945, making them impassable.

The restoration of three of the four road bridges to their pre-war form and thus the connection between the old and new town was completed in 1949. Only the Carola Bridge over the river, which had been completely destroyed, was not initially rebuilt. On March 7, 1952, the last remaining steel girder over the Elbe was blown up (Fig. 1, [4]).

Fig. 1: Blasting of the last remaining steel girder of the Carola Bridge on March 7, 1952 (film sequences from [4]) - Photo: SLUB Dresden, Filmverband Sachsen (Dresden), Hirsch Film (Dresden)

The steadily increasing volume of traffic brought the Carola Bridge back into focus and it became part of the plans for the redesign and traffic development of Dresden's city center after the Second World War [5]. An almost 2.8 km long north-south connection was planned from Platz der Einheit (now Albertplatz) in Dresden's Neustadt district across the Elbe to Wiener Platz at the main railway station. Intensive studies began in 1962. The preferred variant envisaged two directional lanes separated by a grass verge. The streetcar tracks were to be positioned laterally towards the city center [6]. It is interesting to note that similar large-scale redesign plans for Dresden's city center had already been drawn up during the Nazi era in 1938/39 and again in the early 1950s, although they were not widely discussed in the following years [7, 8].

Design competition for a new Elbe crossing [5]

The new bridge had to meet high design standards, particularly due to its immediate proximity to the Baroque architecture on the Old Town side and to the monumental neo-Renaissance and neo-Baroque government buildings built at the end of the 19th/beginning of the 20th century on the New Town side [9], which characterize the city skyline at the bridge site. The launch of a competition for the bridge design was a novelty for the GDR. The main reason for this was that it was deemed impossible to commission several design companies to work on the designs in parallel. Instead, several collectives were asked to develop bridge designs voluntarily and in their free time.

A total of eleven teams took part and designed nine prestressed concrete and four steel solutions with main spans of up to 180 meters. In accordance with the invitation to tender, they were all beam bridges. Overhead supporting structures had been discussed in advance, but were rejected due to the urban planning situation. There were two variants with parallel chord superstructures. All others had one or two haunches above the piers of the power opening.

The main evaluation criteria included the external form (including harmony in the division of the field, cove height, view of the Brühl's Terrace on the Old Town side), the costs and the use of materials. The winning design came from the Dresden collective Thürmer, Spoelgen (Fig. 2). This was also the only collective to submit several - namely a total of three - variants. In [5], the report is quoted as follows: "The idea of spanning the Elbe river with a bridge that is greatly increased in height on the Neustädter Strompfeiler and then narrows accordingly on both banks with a curved lower edge is particularly commendable. This gives the bridge its own character, which is in harmony with the cityscape and the curvature of the river. The resulting asymmetrical solution corresponds to the different proportions of the two banks."

Fig. 2: Excerpt from the submitted competition documents of the later winning design, the existing foundations are shown as dashed lines - drawing: from [5]

The supporting structure [10] Overview

Based on the competition design, the new bridge over the Elbe was planned by VEB Entwurfs- und Ingenieurbüro des Straßenwesens (EIBS), Dresden branch, under the leadership of Eckhart Thürmer. With a main span of 120 m, the Dr. Rudolf Friedrichs Bridge, which was completed on June 10, 1971 and opened to traffic on July 3 of the same year, was the longest-span prestressed concrete bridge in the GDR [11]. It was renamed Carola Bridge in 1992 (e.g. [12, 2]). This name is also used in the following.

Fig. 3: Longitudinal section and ground plan of the Carola Bridge - drawing: from [10]

The structure, consisting of five spans with very different individual support widths between 44 and 120 m, is approx. 400 m long, 32 m wide and has three individual, independent superstructures in the form of single-cell prestressed box girders (Fig. 3), which are referred to as bridge spans a to c. With the chosen single-span geometry, it was possible to meet the strict requirements for the maximum desired gradient height and the required clearance height in the area of the navigation opening. At the same time, the bridge adapts very well to the asymmetrical topology at the bridge site and to the dominant buildings on both banks. The fixed point is located at the pier in axis D, where the superstructure also has the maximum construction height of 5.2 m. It tapers to 1.6 and 1.8 m towards the abutments. The static calculation had revealed strongly varying moments in a continuous solution, which is why the superstructure was subdivided with a total of three joints in order to be able to realize the planned geometry. This resulted in a two-span girder in axes A-C with a 12 m long cantilever in the direction of the Elbe (joint I). A single-span girder with two cantilever arms (44 m long in the direction of the Elbe (joint II) and 10 m long in the direction of Neustadt (joint III)) lies on supports D and E. The 64 m long current suspension girder spans between the two cantilever girders. A further 48 m long suspension girder was designed in the edge span on the Neustadt side.

The structure was designed for bridge class 60 [5]. Trains a and b each crossed two-track directional lanes, train c a two-track tramway route. Trains a and c each had a 3.2 m wide walkway for pedestrians and cyclists on their outer sides [13]. Pipes for district heating, gas, electricity and water were also housed in the box girders.

Superstructure in longitudinal direction

Fig. 4: Neustadt cantilever girder train b; moment curve for various load situations (top), total lines of the BSG 100 bundle tendons (middle) and number of web tendons (bottom) - drawing: from [10], mod.The structural calculations were carried out partly manually and partly on electronic computer systems already available at the time. This made it easier to take into account the constantly changing cross-section values and the effects of creep, shrinkage and changing temperatures. Comparative calculations were carried out for special problems, e.g. at the Institute for Lightweight Construction Dresden or at the TU Dresden.

Finding a sensible tendon layout proved to be difficult. Figure 4 gives an impression, showing the moment curves for different structural conditions. For example, positive bending moments prevail in 70 % of span D-E in the construction state without suspension girders, whereas negative moments clearly predominate in the final state. This was exacerbated by significant creep and shrinkage deformations during the construction period. To solve this problem, three different types of tendons were used: permanent tendons for the dead load, installation tendons for the construction stage and subsequent tendons that were activated for the final stage. In the construction stage, the current-side cantilever arm (joint II) of the first constructed tension c was ballasted using chippings, which were filled into the box girder. This very complex procedure was later replaced for the two other trains a and b by the use of assembly bracing. The movement of the cantilever end was deliberately influenced by the ballasting or bracing. The structural calculations for the various construction stages showed vertical movements at the end of the cantilever arm of between 24.9 cm upwards and 32.5 cm downwards. For the final state, movements of between 14.5 cm upwards and 23.3 cm downwards were still predicted.

Fig. 5: Arrangement of tendons and slack compression reinforcement (left) and stress distribution due to different load combinations (right) in the support cross-section axis D, tension b - drawing: from [10], mod.

The extraordinarily dense longitudinal reinforcement in the support cross-section axis D - in the upper part the tension reinforcement and in the lower part the slack reinforcement - is shown in Fig. 5, left, as an example for tension b. On the right is the stress curve over the cross-section height for various load cases. Under dead load, prestressing (including creep and shrinkage) and traffic, tensile stresses are permitted to a small extent at the top of the cross-section, which corresponds to limited prestressing. A total of 140 bonded tendons (BSG) 100 were installed in the support cross-section D in span b, and even 40 to 50 more tendons of the same size in the neighboring spans a and c. The 100 stands for the nominal tensioning force per tendon in Mp, which corresponds to approx. 1 MN. The tendon lengths were adapted to the moment curve. Tendons that did not extend over the entire length were anchored in brackets in the webs as well as the floor slab and deck slab (Fig. 4 below and Fig. 6).

Massive slack compression reinforcement was installed in ten layers in the floor slab of the axis D support area to ensure the load-bearing capacity in the compression zone of the cross-section. The degree of reinforcement here was 4 % and thus reached that of reinforced concrete columns. Stirrups were used to prevent buckling.

Superstructure in transverse direction

In the transverse direction, the box girders were only loosely reinforced. Particular attention was paid to the internal forces resulting from changing temperatures. Temperature differences were caused on the one hand by high daytime and low night-time temperatures, and on the other hand by the operation of the district heating pipes, which were poorly insulated inside the bridge. In turn, internal forces in the cross-section resulted from restrained deformation of the box girder.

Cross plates were arranged above the supports. These were perforated by various openings for access and cable routing (Fig. 7). Truss and frame models were used for their calculation.

Gerber joints made of cast steel

The three joints in each of the superstructures are unique. Gerber joints were usually designed as concrete brackets with elastomer bearings. For the Dresden Bridge, a special proposal (so-called new proposal) was submitted for "steel joints for solid bridges" [14], according to which the joints were then also executed. The steel castings were attached to the end faces of the webs of the cantilever end and the suspension girder after prestressing the continuous web tendons and tensioned to the anchored tendons with several coupling bolts (Fig. 8).

In [14], in addition to the design and tensioning process, a measuring arrangement for checking the applied forces was also described in outline. In the final state, the joints were not visible from the outside, but were covered by so-called "concrete aprons", which were subsequently concreted on.

Fig. 8: Gerber joints of the Carola Bridge, left: Top view and elevation, top right: Detail of the cast steel web connection, bottom right: Part of joint I, train c, during demolition - drawings: from [10], mod.; photo: Silke Scheerer

Cross connection

Fig. 9: Cross connection between trains a and b - Photo: Silke ScheererThe three individual superstructures are connected to each other by a crossbeam at the level of joint II (Fig. 9). The design as a transverse joint allows for longitudinal and transverse movements of the superstructures as well as twisting. During construction, the transverse connection made it possible to compensate for height differences between the superstructures, which were built one after the other, by means of vertical clamping presses. In the final state, the task of the transverse connection was to compensate for differences in deflection of the three very slender, separate box girders due to shrinkage and creep, temperature and traffic, as well as to distribute loads transversely and thus force the neighboring superstructures to participate in the load transfer.

Construction [15] Construction sequence

In 1966, work began on removing 15,000^m3 of rubble and the remaining parts of the old Carola Bridge. The Elbe divided the bridge construction site into two parts, with most of the construction site equipment being placed on the Old Town side. People were initially transported by motorboat.

The existing foundations were used as far as possible for the construction of the abutments, piers and supports. Work began with the section between axes A to C on the Old Town side. Once the substructures had been constructed, a falsework was erected here and the first part of the superstructure was constructed. At the same time, the old pier 3 in the stream was removed. It was known from eyewitnesses that two aerial bombs had been dropped into the Elbe in 1945 in the area of the planned new bridge construction [16]. One of these bombs was found and recovered during the demolition of the pier in May 1967. The second bomb was only discovered and defused in January 2025 during demolition work on the collapsed bridge c [17]. Two further unexploded bombs were recovered on 27 and 28 January 2025.

The substructures were then constructed in the remaining axes and the scaffolding for the Neustadt cantilever girder was erected. Between the first two construction phases, a lightweight cable structure was erected as a temporary bridge [18]. This was used to transport people and the bundle tendons produced on the Old Town side. It also carried the pressure pipes for the fresh concrete produced on the left bank. The suspended girders formed the 3rd and 4th construction phase.

The falsework was designed to be transversely displaceable to allow multiple use. The different curvature of the three superstructures in plan was a major challenge. In the river area, the falsework was erected from piling barges, one of which also carried a slewing tower crane (Fig. 10). The suspension beams were constructed 20 cm above the nominal position. After prestressing the suspension girders, the cast steel components of the Gerber joints were tensioned at the respective end faces. The clamping force was determined in selected bolts using a micrometer gauge. Finally, the suspension beams were lowered into their final position using presses.

Fig. 10: Assembly of the falsework over the Elbe - Photo: taken from [15]

Handling the prestressing steel

Oil-tempered prestressing wires St 140/160 from Hennigsdorf were used as prestressing reinforcement [19, 20]. The nominal area was 50 ^mm2. In the course of sampling after the collapse of train c, wires with 40 ^mm2 were also found in train b.

The bundle tendons (BSG) were prefabricated on site. The steel bearing was of solid construction in order to meet the normative requirements for corrosion protection in accordance with [21]. The warehouse could be heated. The humidity was monitored by measurement.

First, 24 individual wires were cut to length for each BSG 100. The cladding tube was then pulled onto the steel bundle using winches. For lengths over 50 m, this was done in sections and the ducts were overlapped accordingly. The longest tendons measured 147 m (cantilever beam on the Neustadt side). The anchoring components (cone, sleeve, duct extension) were completed either in the tensioning shed or on site at the installation location.

In order to ensure the best possible corrosion protection, various measures were used, from the prefabrication of the BSG to the pressing of the bundles. In addition to the aforementioned temperature control and moisture monitoring of the tensioning shed, the BSGs were intensively flushed with cold air via the injection and ventilation openings at daily intervals between concreting and prestressing. The main purpose of this was to remove any condensation water that accumulated. No rust formation was visually detected during the inspections. According to witness reports, the process was largely controlled until the tendons were installed in the formwork. From then on, however, the corrosion protection measures are only sparsely documented.

Prestressing began approximately six days after completion of the concreting work for each component. For the suspended beams, partial prestressing was carried out on the fourth day to prevent shrinkage cracks, and 21 days after the end of concreting, losses in tensioning force due to creep and shrinkage were compensated for by post-tensioning. If there is a deviation from the normative specifications in accordance with GBL I Part II No. 84 of 1967 "Directive on corrosion protection for prestressed concrete" [21] - which states in §7 Paragraph (4) that St 140/160 must not remain unpressed in prestressing channels for longer than ten days - special measures must be taken in accordance with Paragraph (7) [21]. Listed here are inert gas in the cladding tube, protective liquids or hot air. In any case, these measures require the approval of the Office for Metrology and Product Testing or the state building inspectorate of the Ministry of Construction at the time.

Special features of concreting

Fig. 11: Sampling point for floor slab train b, Neustadt side, in the longitudinal direction of the bridge (horizontal in the picture) two slack steels, a pipe set in concrete and an open tendon can be seen - Photo: Silke ScheererTheCarola Bridge was designed with B450 standard concrete. To ensure consistent concrete quality and appearance, the aggregates, for example, were stored in storage boxes with a total capacity of 1,000^m3. In order to keep the moisture as constant as possible, the boxes were equipped with drainage. As there was little experience with the pumpability and shrinkage behavior of the B450, extensive technology tests were carried out in advance.

In winter, the aggregates were preheated with steam pipes. Hot water was also used for mixing. The mixing plant was also enclosed so that concrete could be produced even at temperatures as low as -15 °C [22]. In very cold conditions, warm air was introduced into the covered formwork before concreting to prevent the fresh concrete from cooling too quickly.

To avoid damage in the construction joint between the floor slab and the web of the cantilever beam in axis D due to the expected considerable hydration heat development, pipes were embedded in the concrete for cooling (Fig. 11), which were later pressed out like the cladding pipes of the BSG.

The article is based on a presentation at the 3-country corrosion conference in Dübendorf near Zurich at the end of March.

The second and final part will be published in the next issue of Galvanotechnik.

Literature
[1] Oelsner, N.: Die Dresdner Elbbrücke im Mittelalter und in der frühen Neuzeit. In: Dresdner Geschichtsverein e.V. (ed.): Dresdner Elbbrücken in acht Jahrhunderten, published in: Dresdner Hefte 26 (2008) 94, Dresden: Michael Sandstein, Grafischer Betrieb und Verlagsgesellschaft mbH, 2008, pp. 5-14.
[2] Dresdner Geschichtsverein e.V. (ed.): Dresdner Elbbrücken in acht Jahrhunderten, published in: Dresdner Hefte 26 (2008) 94, Dresden: Michael Sandstein, Grafischer Betrieb und Verlagsgesellschaft mbH, 2008.
[3] List of Elbe crossings in Dresden at: https://de.wikipedia.org/wiki/Liste_der_Elbquerungen_in_Dresden incl. further links ibid.
[4] Hirsch, E. (director): Carola-Brücke Dresden, Sprengung des alten Brückenträgers am 7. März 1952 [Archive title]. BR Germany, short documentary film, 1952 - available at: https://www.filmportal.de/node/1726836/video/1731223.
[5] Slavik, V.: Study competition for the new Elbe bridge in Dresden. Bauplanung - Bautechnik 18 (1964) 3, pp: 124-135.
[6] Peschel, R.: Verkehrsplanerische und verkehrstechnische Gestaltung der Nord-Süd-Verbindung in Dresden. In: Bau der Nord-Süd-Verbindung in Dresden : Erinnerungsschrift, Dresden, 1971, pp. 10-14.
[7] Fiedler, E.: Straßenbrücken über die Elbe in Deutschland : eine Darstellung der historischen Entwicklung dieser Brücken. Dresden, 2005.
[8] Kantschew, T.: Das geplante "Gauforum Dresden" - Werkzeug zur Massenmanipulation - Gigantomanie des deutschen Faschismus. Published by Das neue Dresden - https://www.das-neue-dresden.de/gauforum.html, 2005 incl. later additions.
[9] Wikipedia entries on the development on the Neustadt side: https://de.wikipedia.org/wiki/S" target="_blank">https://de.wikipedia.org/wiki/Sächsisches_Staatsministerium_der_Finanzenand https://de.wikipedia.org/wiki/S" target="_blank">https://de.wikipedia.org/wiki/Sächsische_Staatskanzlei#Gebäude.
[10] Berger, R.; Franke, M.; Thürmer, E.: Projektierung der Dr.-Rudolfs-Friedrich-Brücke. DIE STRASSE 11 (1971) 6, PP. 266-277.
[11] Schlimper, H.: Ceremonial address. In: Bau der Nord-Süd-Verbindung in Dresden : Erinnerungsschrift, Dresden, 1971, pp. 1-3.
[12] Schleicher, C.: Long-term investigations on the Carolabrücke Dresden. Structural Engineering 71 (1994) 1, pp. 15-22.
[13] Kriesche, K.: Design and construction specifications for the Dr. Rudolf Friedrich Bridge. DIE STRASSE 11 (1971) 6, PP. 264-277.
[14] Gremler, A.; Fischer, K.-H.; Thürmer, E.: NV 84/69 "Steel joints for solid bridges". Submitted to the Büro für die Neuererbewegung (BfN) on 15.12.1969, accepted on 20.02.1970.
[15] Fleischer, H.; Göbel, W.; Haffner, P.; Kluge, P.; Kamjunke, K.-H.; Riedrich, W.; Römhild, H.; Steglich, K.: Bauausführung der Dr.-Rudolfs-Friedrich-Brücke. DIE STRASSE 11 (1971) 6, PP. 278-290.
[16] Contemporary witness interview by Jakob Vogt (IMB, TU Dresden) with Dipl.-Ing. Hilmar Uhlich, construction manager of the Carola Bridge from January 1967 to September 1968, in October 2024.
[17] Endt, C.; Schloms, M.; Kuhn, H.; Berndt, H.; Anders, F.; Siebert, P.; Schneider, A.; Just, J.: World War bomb at the Carola Bridge: Dresden in a state of emergency for more than a day. Sächsische Zeitung online, 09.01.2025.
[18] Fischer, K.-H.; Göbel, W.: Cable construction as a temporary bridge. DIE STRASSE 10 (1970) 6, PP. 334-337.
[19] TGL 101-036, Sheet 1: Prestressing steel St 140/160 oil quenched and tempered Stahlmarken Technische Lieferbedingungen. 01.08.1966-31.12.1968.
[20] TGL 101-036, Sheet 2: Prestressing steel St 140/160 oil quenched and tempered oval ribbed dimensions. 01.08.1966-31.12.1968.
[21] Order on corrosion protection for prestressed concrete dated August 19, 1967, published in: Law Gazette of the German Democratic Republic, Part II No. 84, date of issue: 08.09.1967, pp. 588-592.
[22] Riedrich, W.: Winter construction problems during the construction of the Dr. Rudolf Friedrich Bridge. Construction planning - construction technology 25 (1971) 9, pp. 453-455.