Monitoring and Modelling of the Abiskojåkka Bridge

(1)

Monitoring and Modelling of Abiskojåkka Bridge

Kasper Furenstam

Civil Engineering, master's level 2020

Luleå University of Technology

Department of Civil, Environmental and Natural Resources Engineering

(2)

Kasper Furenstam

MONITORING AND MODELLING OF THE ABISKJÅKKA BRIDGE

Master thesis Väg- och vattenbyggnad, inriktning Konstruktion, 30,0 hp, X7010B

Structural and Fire Engineering

Department of Civil, Environmental and Natural Resources Engineering Luleå University of Technology

971 87 Luleå

(3)

(4)

Abstract

The infrastructure of today is getting older and problems caused by deterioration over time is affecting the service life of these structures. In Sweden most of the existing bridges were constructed 60 to 70 years ago, rising the need to determine the state of health of the bridges as the maintenance costs will increase heavily. Part of the above-mentioned cut of the bridges owned by Trafikverket (The Swedish Transport Administration) that was constructed 60 – 70 years ago are pre-stressed concrete bridges. Pre-stressing of concrete structures is today a commonly used technology that utilizes the beneficial characteristic of concrete, the compressive strength, to a further extent than reinforced concrete.

This report will focus on the problems with pre-stressed concrete bridges and particularly on the thermal effects on the Abiskojokk railway bridge located in the northern part of Sweden. The pre-stressed box girder bridge spans in total 86 m in three lengths of 30 m, 35 m respectively 21 m starting from the east abutment and is part of the Iron-Ore Line starting in Kiruna and ending in Narvik, Norway. In an ocular(särskild) inspection of the bridge carried out the 18^th of August in 2016 several crack patterns were mapped on the inside of the box girder along with some cracks about the top of the first column support starting from the east abutment. This thesis is focusing on the cracks that was mapped along the tendon positions on the inside of the box girder in the first span starting from the east abutment. The hypothesis is that the cracks are caused by temperature loads and normal forces obtained from the tendons at the thickening of the cross- sections. The research questions are; what the monitoring program shows and if it is possible to prove the hypothesis by using of a FE-model considering the gravity loads, temperature loads and the pre-stressing.

In order to determine the cause of the cracks on the inside of the box girder and investigate the behaviour of the bridge a monitoring program was installed, measuring the crack development over time along with the acceleration and temperatures of the bridge. Example data from the program were later used to analyse the behaviour of the bridge.

The results from the temperature data shows that the bridge has a slowness to temperature changes outside. This gives rise to temperature gradient acting over the bridge parts that may contribute to crack propagation. It also showed that the temperature correlates well with the strain of the cracks. The LVDT’s showed that the largest crack openings during train loading occurred in the second span of the bridge. The data also showed relatively large and unexpected negative peaks during the train loads. The strain gauges also show that the largest strain is occurring in the second span of the bridge. The crack envelopment during a train loads are more expected here and may prove that the negative peaks from the LVDT’s and accelerometers are caused by vibrations. The accelerometers showed that the largest transversal accelerations take place in the first and third span. This may be due to more restricted supports conditions at the column supports than the abutments. The accelerometer also showed correlating negative peaks with the LVDT’s that may be caused by vibrations in the bridge.

The conclusion from the monitoring program so far is that is not possible to prove the cause of the cracks so far, but it may be in the future.

The results from the non-linear FE-model showed that the thermal action of the Eurocode gradient was not enough to crack the concrete along with the pre-stressing load. However, the effects of the hypothesis were proven right.

…

Keyword:

Pre-stressed bridge, Abiskojokk, Monitoring, thermal effects, FE modelling, ATENA

(5)

(6)

Sammanfattning

Dagens infrastruktur blir allt äldre och med tiden sker en försämring av byggnadsverkens prestanda på grund av olika nedbrytningsprocesser, vilket påverkar livslängden på byggnadsverken och som därmed spelar en viktig roll vad gäller byggandet av ett hållbart samhälle. Den största delen av brobeståndet i Sverige som förvaltas av Trafikverket byggdes för 60–70 år sedan och kommer i framtiden innebära eventuellt stora kostnader i takt med att broarna börjar uttjäna respektives livslängs. Denna stora broandel lyfter därmed, över tid, behovet för tillståndsbedömning i och med de stora hållbarhetsvinster som finns att tillgå vid en eventuell förlängning av broarnas livslängd.

En stor andel av den ovan nämnda broandelen, som ägs och förvaltas av Trafikverket och byggdes 60–70 år sedan, är förspända betongbroar. Förspänning av betongkonstruktioner är idag en vanligen använd teknik som nyttjar betongens starka tryckhållfasthet i högre utsträckning än vad ordinära armerade betongkonstruktioner gör. Användandet av dragförband gör att en större andel av spänningsfördelningen över tvärsnittet blir i tryckspänning. Det finns dock problem med användning av tekniken. Krympning och krypning av betongen över lång tid leder till att spännkraft går förlorad och att dragspänningen i tvärsnittet ökar. Fler fall av temperaturrelaterade spänningsproblem har också rapporterats genom åren.

Denna rapport kommer att fokusera på problematik relaterat till spännarmerade broar och speciellt de problem som har upptäckts på bron över Abiskojokk i den norra delen av Sverige.

Den förspända brolådan spänner totalt över 86 m i längder av 30 m, 35 m repektive 21 m sett från det östra landstödet och utgör en del av Malmbanan som sträcker sig mellan Kiruna och Narvik (Norge). Vid en särskild besiktning utförd den 18:e augusti 2016 dokumenterades ett flertal sprickor i bron. Dels längs med insidan av brolådan i longitudinell sträckning och dels i området kring det första pelarstödet sett från det östra landstödet. En bärighetsberäkning utfördes också för bron 2016 i och med att axellasterna skulle ökas.

Den hypotes som har ställts upp i arbetet är att sprickorna beror på temperatur och normalkrafter från spännkablarna. Frågeställningen är om det går att bevisa hypotesen med hjälp av ett tillståndsbevakningsprogram, vilka slutsatser som kan dras från datan och om det med hjälp av en FEM-simulering som beaktar egentyngder, spännkrafter och temperatureffekter går att bevisa hypotesen. Tillståndsbevakningsprogrammet innefattar mätningar av sprickdeformation, spricktöjningar, acceleration och temperatur som mäts över tid. Arbetet fokuserar på sprickorna som återfinns på insidan av brolådan. För att sammanställa resultaten i projektet var målet att bygga upp en BIM där alla resultat skulle samlas.

Mätutrustningen installerades under två fältbesök som utfördes i februari och i maj. Under det första fältbesöket installerades totalt fyra sprickgivare, en accelerometer, sex termoelement, en thermocentral, två noder och en basstation. Sprickgivarna (LVDT’s) placerades parvis över sprickorna 115 och 106 som återfinns i den rapport som gjordes i samband med en särskilda inspektion 2016. Accelerometern placerades nära det första spannets mitt och termoelementen placerades i de två väggarna för att kunna mäta den gradient som uppkommer. Termoelementen är kopplade till en termocentral som sänder datan vidare till basstationen. Från givarna (sprickgivarna, accelerometer och termelement) sänds datan trådlöst över till en Gateway i bastationen som via ett modem och en antenn kopplar upp mot GSM-nätet och för datan vidare till molnet där datan kan hämtas. Till basstationen leds 230 V in som konverteras till 12 V som systemet drivs av. Som en reserv finns ett batteri som laddas kontinuerligt och används om ett strömavbrott skulle ske. Från basstationen leds 12 V ut till termocentralen, accelerometern och noderna till sprickgivarna.

Under det andra fältbesöket som utfördes i början av maj installerades totalt sex sprickgivare, två eccelerometrar, två noder, två strömdosor och en riktantenn. I det andra spannet installerades

(7)

tre sprickgivare, två över sprickorna 112 och 113 och en över 103 på den norra väggen.

Accelerometern monterades på den södra väggen och noden för sprickgivarna och en eldosa likaså. I det tredje spannet installerades sprickgivare över sprickorna 110 och 102. Två av dem på den södra väggen/golvet där även noden, accelerometern och strömdosan installerades. För att förbättra signalen mot Gatewayen i basstationen installerades en riktantenn i taket av det tredje spannet också. Vid ett senare tillfälle installerades också töjningsgivare på varje siaoch i varje spann.

Exempeldata från tillståndsbevakningsprogrammet har hämtats för att undersöka om det finns samband mellan olika effekter som temperatur, spännkablarnas placering etc. Datan omvandlades till nya enheter och tidsangivelser innan den illustrerades i grafer.

Vidare har en FE-model av bron gjorts för att undersöka temperaturens och spännkablarnas effekter på sprickor i regionen där sprickor har uppkommit i det första spannet. Det första steget i detta arbete var att rita upp brons alla armeringsjärn vilket gjordes i AutoCad. Därefter importerades denna Cad-fil till FEM programmet ATENA. I detta program tilldelades de olika materialen egenskaper utifrån de ritningar som har funnits att tillgå i BaTMan samt den bärighetsberäkning som utfördes 2016 på uppdrag av Trafikverket. Modellen belstades av egentyngder, förspänningskrafter och temperaturlaster enligt rådande normer. På grund av modellens storlek och utformning gjordes en förenkling där kantbalkarna togs bort och där bron delades i mitten av det andra spannet för att kunna köras. Modellen tilldelades en icke-linjär zon där sprickor har uppkommit i den verkliga bron. I detta område studerades också effekten från temperaturlasterna i nivå med spännkablarna på den norra väggen i brolådan.

För att validera den icke-linjära modellen gjordes också linjär FE-model i programmet AxisVM. Denna model gjordes på grund av begränsningar i programvaran rak vilket är en förenkling av brongeometrin. Spännkablarnas placering korrigerades därefter och samma materialparamterar som för den icke-linjär modellen tilldelades bron. Modellen består av totalt sex tvärsnitt som är fördelade enligt ritningarna av bron och överbyggnaden har gjorts med rib- element. Denna model tillsammans med värden från bärighetsberäkningen och Trafikverkets rekommendationer har använts för att validera den icke-linjära modellen.

Temperaturdatan som har hämtats visar att temperaturen vid mätstationen för SMHI skiljer sig från den som har uppmäts i brolådan. Detta är väntat och visar på att brons insida har en tröghet gällande temperaturförändringar som leder till temperatur gradienter över brolådans delar. Detta visar på att det finns ett intresse för att undersöka temperaturens effekter på brolådan. Den mest intressanta perioden sker dock under hösten då positiva temperaturgradienter förväntas uppstå som bör leda till ökade dragspänningar i brolådan. Vidare visar temperaturmätningarna på att temperaturskilknader över broväggarnas tjocklek förekommer från mätningarna med termoelementen.

Sprickgivarna visar att de störst spickrörelserna sker i det mittersta spannet av bron under tåglast. Vidare visar sprickgivarna på oväntade negative dalar vilket kan tyda på vibrationer i bron som påverkat sprickgivarna. Dessa återfinns tydligast i det tredje och andra spannet.

Töjningsgivarna visar frånsett en töjningsgivare att de strörst töjningarna sker i det mittersta spannet i lekhet med sprickgivarna. Töjningsgivarna visar tydliga öppningar av sprickorna under tåglaster och inga negativa dalar, vilket mest troligt bevisar att de negativa utstickande dalarna beror på vibrationer i bron.

Accelerometrarna visar att de största accelerationern i trnasversal riktning sker i det första och det tredje spannet av bron. Detta kan bero på att pelarstöden begränsar rörelsen i transversal riktning mer än vad landstödden gör. Vidare visar accelerometerarna negativa utstickande värden som sammanfaller med det negativa utstickande dalar som återfinns i datan från sprickgivarna.

Vibrationerna, som detta mest troligt beror på, fångas av accelerometern varför det bör finnas ett samband i detta.

Som svar gällande frågeställningen har det inte gått att visa vad sprickorna beror på under denna uppsatstid med hjälp av tillståndsövervakning. Det kan dock hända att det går i framtiden.

(8)

Som frågeställning i FEM-delen ställdes frågan om det var möjligt med hjälp av en icke-linjär model simulera det funna sprickmönstret på insidan av brolådans första spann. Resultatet visade att sprickmönstret inte gick att återfinna i brolådan. Däremote visade resultaten att det finns ett klart samband mellan en avkylning av brons yttersida och ökade töjningar i brolådans insida.

Högre förspänning resulterade likaså i större töjningar i brolådans insida. För att förbättra modellen har rekommendationer framförts. Framförallt gällande en större model som beaktar brons hela geometri samt tåglasten på bron. De ritningar som har gjorts av bron i IGES-format kan vara väldigt hjälpsamma vid en sådan förbättring.

För att redovisa och samla resultaten från tillståndsbevakningsprogrammet och FEM har en BIM påbörjats och scanning av brons insida har gjorts för att redovisa datan på ett pedagogiskt sätt. Denna scan kan sedan importeras in i en framtida BIM för att snyggt kunna illustrera utvecklingen i arbetet över tid. En påbörjad BIM har gjorts i Revit.

Slutligen konstateras det att detta arbete inte är slutfört gällande orsaken till sprickorna och att mer arbete inom området måste ske varför fler områden som är lämpliga för examensarbeten har förslagits. Bland annat non-destructive testing och utveckling av BIM-modellen, big data analys av datan , förbättrad FE-modell och scanning av bron med hjälp av drönare.

…

Nyckelord:

spännarmerad bro, Abiskojokk, tillståndsbevakning, FE-analys, temperatureffekter, ATENA

(9)

Preface

This master thesis project at Luleå University of Technology (LTU) was initiated by Professor Björn Täljsten and Mr. Ibrahim Coric from the Swedish Transport Adminitration, Trafikverket.

Along with Professor Täljsten and Mr. Coric, Technical Doctor Cosmin Popescu has been the supervisor during the project.

I am very thankful for the support I have received from Carlos Lindau at LISAB during the in- stallation of the monitoring equipment and collecting the data from the system. I also want to thank my brother Elias and Sofie for their help with handling of the data and Adam Söderström and Carlos Lindau for their support with the data representation.

Moreover, I want to thank Jani Mukkavaara and Adam Oskarsson for their support with the Revit model in the project. I would also like to thank Dobromil Pryl and the Cervenka team for the countless answers to my questions and Erika Bokor along with the Axis team for the free student license that was used in the project. I also want to thank Herlander Sapage for the support with the license servers at the university.

I want to express my sincere gratitude to doctor Popescu for the guidance and support through the thesis and to Professor Täljsten and Mr. Coric for initializing the project. I also want to thank Professor Lennart Elfgren for the helpful guidance, comments and for the chance to explore this field of interest.

Lastly, I want to express gratitude to my mother for the support through out the years.

Luleå, June 2020 Kasper Furenstam

(10)

List of Abbreviations

ATENA: FEM Software AxisVM: FEM Software

BIM: Building Information Model

CFRP: Carbon Fibre Reinforced Concrete FE: Finite Element

FEA: Finite Element Analysis FEM: Finite Element Method GiD: Pre- and post-processor GSM: Global System for Mobile

LVDT: Linear Variable Differential Transformers NDT: Non-destructive testing

PC: Pre-Stressed Concrete RC: Reinforced Concrete SLS: Service Limit State

TRVK: Trafikverkets råd och krav ULS: Ultimate Limit State

(11)

List of Notations

𝑨_{𝒆𝒍𝒐𝒏𝒈𝒂𝒕𝒆𝒅}: Reduced area because of elongation 𝑨_𝒊: Area in a cracked condition

𝑨_𝒑: Area of one tendon 𝑨_𝟎: Area uncracked 𝑪_𝑻: Heat capacity

𝑬_𝒄𝒅: Concrete design elastic modulus 𝑬_𝒄𝒎: Concrete mean elastic modulus 𝑬_𝒑: Young’s modulus for the tendon steel 𝑬_𝒕: Tensional Young’s modulus

𝑱_𝑻: Heat flux

𝑲^𝑺: Secant stiffness matrix 𝑲^𝟎: Initial stiffness matrix

𝒂_𝒌: Linear combination coefficient 𝒇_𝒄𝒅: Concrete design compressive strength

𝒇_𝒄𝒌: Concrete characteristic compressive cylinder strength after 28 days 𝒇_𝒄𝒎: Concrete mean cylindrical compressive strength

𝒇_𝒄𝒕𝒅: Concrete design tensile strength 𝒇_𝒄𝒕𝒎: Concrete mean tensile strength 𝒇_𝒄𝒖: Concrete Compressive Strength 𝒇_𝒑𝒌: Charcteristic tensile strength 𝒇_𝒔𝒖: Steel characteristic ultimate strength 𝒇_𝒔𝒚: Steel characteristic yield strength 𝒇_𝒕: Maximum tensile stress

𝒇_𝒖𝒅: Steel design ultimate strength 𝒇_𝒚𝒅: Steel design yield strength 𝒒_𝒃𝒂𝒍: The ballast load

𝒘_𝒄: Crack band width

𝜸_𝒃: The heaviness of the ballast 𝜸_𝒄: The heaviness of the concrete 𝜸_𝒏: Partial safety factor

𝜺′^𝒑𝒍: Elastic strain rate 𝜺_𝒄: Compressive strain

𝜺_𝒄𝟏: Peak compressive strain at peak compressive stress 𝜺_𝒇: Microcrack strian

𝜺_𝒑: The strain of the prestressed tendon

𝜺_{𝒑𝟏𝟏𝟎}: The strain of the prestressed tendon with 10 % extra strain.

𝝈_𝒃𝒂𝒍: The stress from the ballast 𝝈_𝒄: Compressive stress

𝝈_𝒆𝒃: The stress from the edge beam A: Area of chord

𝑫: Diagonal matrix 𝑮: Gauge factor

𝑮: Plastic potential function 𝑯: Inflow of heat

𝑲: Stiffness matrix

𝑳: Lower/higher triangle matrix 𝑸: Heat supply

(12)

𝑹: Nodal forces 𝑻: Temperature 𝒂: Displcament vector 𝒅: Damage parameter 𝒇: Element force vector

𝒌: Factor used to obtain the stress strain relation of concret 𝒌: Scalar

𝒌: Thermal conductivity 𝒍: Length

𝒒: Deformation vector 𝒒: Heat flux

𝒒: Load vector

𝒖: Nodal displacement 𝒗: Poison’s ratio 𝒗: Weight function 𝒙: Length of the chord

𝜼: Relation between the compressive strain and the peak compressive strain at peak compressive stress

𝜿: Hardening parameter 𝝀: Extra degree of freedom

𝝁: Friction coeffecient of the tendon sleeve 𝝆: Resistivity

(13)

1 Introduction

1.1 Background

The increasing traffic loads and speeds leads to challenges in terms of sustainability of today’s infrastructure and the increases are not predicted to stop, but more likely to continue to increase throughout the next decade. In order to build a sustainable future, the infrastructure must be re- tained and used instead of being demolished and rebuilt. Hence, the field of determine the health of structures has great importance, Bergström (2009).

A large part of todays railroad bridges that are over 100 years old are steel structures in Europe.

This is because the infrastrucre needed to be rebuilt fast to meet the high demand after the second world war. The papar do however show that a large part of the railroad bridges not older than 50 year are made of concrete, Dinas A, Nikolaidis et. al (2017). This will likely raise the damand of prolonging the service life of concrete bridges in the future and therefore the use of health monitoring.

Today there are about 16 600 road bridges (Trafikverket (2020)) and 3870 railroad bridged in Sweden(Trafikverket2 (2020)). According to BaTMan (2018) about 700 of these are pre-stressed concrete bridges, neglecting the pedestrian bridges. Hejll (2007) highlights the age of the existing road bridges in Sweden. It shows that most of the bridges were constructed 60 to 70 years ago, which raise the need for maintenece and evan repairs because of the poor maintenance throughout the years.

1.2 Pre-stressed bridges

Täljsten, et al. (2019) states that the reinforced concrete (RC) was first discovered in France in the 1850’s by Josef Monier. To enhance the durability of his pots he fitted an iron mesh into the cast, which later resulted in the first reinforced concrete (RC) bridge. RC is a composite material normally consisting of concrete and reinforcement made from steel, there are however other types of reinforcement made from other materials, e.g. polymer materials.

Concrete is a composite material consisting of cement, water and aggregates. This mix gives the characteristics as an isotropic and heterogeneous material with relatively great compressive strength and low tensile strength. Because of the low strength in tension of the concrete, steel reinforcement is normally embedded into the concrete.

There are different types of RC and PC bridges that often are categorized in many ways. These bridges can be pre-casted or in situ casted, have arches or cables carrying the decks etc. Focusing on the most common type of bridges, the simply supported continuous bridges, there are six common categories according to Pipinato A. (2016). The most common of them, the Solid slab with rectangular cross-section recommended for spans up to 15 m. Voided slab with rectangular cross- section is also a common typ. The voids save self-weight of the superstructure and could therefore span longer distances, recommended distance is 20 – 40 m. Another way to increase the span length of a rectangular slab cross-section is to use a ribbed slab, where smaller internal beams are created within the slab and making the cross-section locally taller. The recommended span length for this cross-section is 20 – 40 m. If the mentioned internal beams are made even taller, they fall into the next categorie, cast-in-sity beam and slab system, with ranging spans from 30 to 50 m.

Another category is the pre-cast girder category, where the girders are pre-casted (often in facto- ries) and then brought to the construction site where they are placed upon the supports. Making the construction process fast. The last category is the box girder bridge that often is used for longer spans that 80 m. The box girder is beneficial for the torsional rigidity that could be crucial for longer spans. To reach the upper limits of recommended span lengths the sections often needs to be pre-stressed which is possible and done in all of the categories.

Pre-stressed concrete (PC) bridges are a commonly used bridge type today. The pre-stressing technique utilizes the mentioned compressive strength of the concrete to handle greater loads.

(17)

The pre-stressing is done using tendons embedded in the section of the structure which is put into tension using hydraulic jacks for example. The tension obtained in the tendons are then kept using anchorages in the concrete section. The tension in the tendons are transferred to the concrete section creating compression. The main advantage, above-mentioned, is that the concrete is less exposed to tension making it possible to build longer spans for example. The main drawbacks that the technique requires high strength concrete and high-tension steel, special equipment like hydraulic jacks, anchorage and highly skilled supervisors and workers.

Prestressing of structures are today widely used in several types of structures, for example buildings, power stations, offshore structures, underground structures, bridges etc. According to Dinges, T. (2009) the concept of PC was firstly seen when P.H. Jackson got the patent to the concept in 1888 in United States. By that time the steel wasn’t good enough to carry out the demands of PC and the technique needed far more work to become applicable. It was not until Eugene Freyssinet in the begining of the 20^th century started using steel with high elastic limit, used more pre-stressing force and high strength concrete the technique went into construction.

Freyssinet started a PC factory in the 1930’s building telegraph poles in concrete pre-stressed using piano wires. After Freyssinet’s discoveries a civil engineer called Gustave Magnel through is exceptional teaching skills, fluent English speaking and great knowledge in the structural engineering field developed learned and developed the PC knowledge from Freyssinet during isolation during World War II. By the time of his isolation he was professor at Ghent university and had access to a laboratory where he studied PC. He had a great advantage by his fluent English and spread his research to the world. This was the beginning of a great period of developing pre- stressing as a concept in the middle of the 20^th century and forth while.

Figure 1 - The Walnut Lane Bridge (Janberg, N. (2009)).

Dinges, T. (2009) describes the first major PC structure in the United States – The Walnut Lane Bridge, see Figure 1. The bridge was built in 1948 and the cross-section consisted of seven T-girders with a continuous upper flange(deck). The PC concept made it possible to build the 49 m long span with a relatively low cost, so low that the offer at first was reject since bids below a calculated bid from a project engineer was automatically rejected at that time. The construction

(18)

of the bridge was a milestone in the development of the concept of PC since the normally con- servative boards accepted the concept being used in a major project.

Figure 2 - The Double Cantilever Method put to use without the use of scaffolding. (Mag- nel, G. (1954))

In 1950 another milestone was achieved in terms of developing the concept of PC – The Dou- ble Cantilever Method, see Figure 2. The concept made it possible to construct large span PC bridges without scaffolding also making it possible to have a faster construction speed since the superstructure is casted in two directions and that this can be done for multiple supports. The concept also made it possible to remain boat traffic beneath the construction for example which may be of great public interest.

In 1964 one of the most famous PC box girder bridges at that time was completed – The Bendorf Bridge. The design won the competition that had the following criteria’s; lowest cost, remain the boat traffic on the river and have a nice design. Using the PC concept along with the Double Cantilever Method and a hinge in the midspan the Bendorf Bridge was constructed.

When longer span lengths are required it is often beneficial to use PC since it utilizes more of the materials being used. It, however, requires highly skilled designers and workers at site hence the risk using the technique. The need for larger scaffolding for PC box girder bridges are also a drawback compared to steel box girder sections that can be manufactured in mechanical work- shops and then welded together at the work site and launched from one side. While a heavy box girder section even may need ground improvements to carry the scaffolding like piles etc. This may be the reason why the former popular technique of PC box girders has decreased in popu- larity of today’s bridge design. There are however many bridges of this type that are getting older why it is important to understand the behaviour of the structure for maintenance etc.

(19)

1.2.1 Abiskojokk bridge

Abiskojokk bridge is a PC box-girder bridge along the Iron-Ore Line north of Kiruna, Sweden, see Figure 3. The bridge spans, in total, 86 m in a curve with the radius of 806 m in three spans of 30 m, 35 m and 21 m respectively from the Abisko side. The construction phase of the bridge started in 1978.

Figure 3 - The Abiskojokk bridge from the pedestrian trail close to the east abutment look- ing towards Riksgränsen.

1.3 Thermal effects

A superstructure like the Abiskojokk bridge is exposed to great climate variations. In the winter there is almost no sun radiation, but it is usually very low. In summer, the daylight is longer with higher temperatures acting on the bridge. This is due to the fact that the bridge is located in the northernmost parts of Sweden, above the arctic circle in fact. Figure 3 shows the north-east side of the bridge. The orientation of the bridge means this side will be most affected by solar radiation in the forenoon and the other in the afternoon.

Because the bridge spans a river and is relatively close located to the big lake, Torneträsk, it is also subjected to humid conditions along with the rain. These aspects may amplify the thermal effects and contribute to the deterioration process of the concrete in a long-term perspective.

An example of a typical thermal crack is when a concrete structure is subjected to sun radiation on one of the sides, normally, causing a positive temperature difference (gradient). Because of the expansion in the high temperature parts of the structure expansion will occur. This may cause cracking if the expansions are restrained, creating internal stresses in the concrete.

1.4 Inspection of bridges

The type of maintenance needed for bridges in Sweden is decided upon by the basis of visual inspections carried out, periodically, every six years from the finalisation of the construction phase of the bridge (Trafikverket. (2018)). According to Hejll (2007), there are several drawbacks with the visual inspections; different inspection professionals notice different problems, degra- dations may occur faster than the time period of 6 years etc.

(20)

The drawbacks give rise to improved ways of inspection like the use of building information models (BIM). The concepts mean that information is stored in a model of the structure and therefore makes it easier to access and visualise in relation to the structure itself. Other technol- ogies may be combined with the models like infrared scanning that results in point clouds with coordinates giving the “real” dimensions of the structure. This may be a good complement to the measurements conducted from the drawings of the structure. The results from the scans may be added as layers in the BIM, then giving an overall view of the structure. One program to build a BIM in is Revit, which has been used in this project.

By the use of Non- destructive testing (NDT) different data can be obtained that usually requires laboratory testing, for example compressive strength of the concrete, location of tendons, internal damages etc.

1.5 Monitoring

Monitoring is an intelligent and powerful tool used in the process of assessments of structures. It gives the ability to display the change of a chosen parameters over time which could be used for multiple options. One example is refining or validating Finite Element Method (FEM) analyses.

Monitoring could take place in cases of assessment events. In the normal event it is part of an assessment of the cause to a problem or evaluation of an improvement of some sort. There are several parameters that could be measured depending on the purpose of the monitoring. Common measurements of interest could be acceleration, crack development, temperature etc.

The monitoring program could be done regarding different extents of measurements. A global program regards the whole structure while a local get more into details. The best practice is usually to use both levels of details combined, complementing each other. However, some measuring devices may be seen as both global and local, for example accelerometers may be giving both local and global information.

1.6 FEM

The best way to study a structure is to conduct a full-scale test, this is often not possible since the structure may be in service or the great expense of such a test. To work around this problem, the use of Finite Element Modelling (FEM) is often used, often much more economically and sus- tainably efficient. There are several commercial software packages to choose from the market.

In this thesis ATENA and AxisVM was used. ATENA science is a more scientifically oriented program more focused on concrete structures. ATENA is able to analyse structures using non- linear properties while AxisVm is a commercial software that is more user friendly and more adopted for design than ATENA. ATENA has won prices regarding shear failure predictions tests and is developed in Czech Republic. Some example of their ongoing research topics are; Effects of underground excavations on damage of masonry buildings for Metrostav in Prha, non-linear analysis of shear concrete anchors for Hilti in Liechtenstein and analysis of concrete tunnel lining under extreme loading conditions for Aegerter&Bisshardt in Switzerland.

AxisVM is developed in Hungary and is adopted to Eurocode and some national standards like DIN for example. The program has been used in bridge projects like; The Upper Liard River Bridge Jacking Canada in 2016 and the reconstruction of Margaret bridge in Hungary. AxisVm do also have the ability run non-linear analyses but not with as many options as ATENA.

The 5.6.1 ATENA-version was used along with the pre- and post-processor GiD 13.0.4. Ax- isVM V4 that was used is a more user-friendly software.

(21)

1.7 Aim and scope

The aim of the project is to investigate the cause of the cracking and evaluate the likelihood of future deterioration. This is part of a larger project and I will focus on investigating the likelihood of temperature effects creating the cracks inside of the webs of the box girder.

My objectives are to start mapping the damages using monitoring and scanning, analyse the temperature effects using FEM and adding the results in a BIM-model.

The results from the continuing research project initiated by Luleå University of Technology and Trafikverket will be collected in a BIM-model, according to Figure 4.

Figure 4 – The planned workflow of the project.

1.8 Hypothesis and research questions The hypothesis of this this thesis was:

The thermal effect along with the radial force acting from the tendons give rise to a negative moment in the walls, which lead to cracking of the walls along the tendons, explained by Figure 5 and Figure 6.

Figure 5 - The normal force coming from the tendons.

BIM

FEM Monitoring

Scanning

Drawings NDT

Material data

Strengthenings

Photogrammetry

(22)

Figure 6 - The resulting deflection and crack oriented along the tendon in the wall of the box girder.

The research questions were:

I. What is the reason behind the cracks appearing in the box girder?

II. Can non-linear FEM simulate the existing crack pattern?

1.9 Limitations

The following limitations was considered:

• Only data sampled from the monitoring done from February to June will be evaluated.

• The cracks inside the first span of the inner wall close to the first column support of the box girder will be examined using non-linear FEM.

• The effects of creep, shrinkage, structural effects (like higher loads), casting defects and combinations of these with the temperature effects have been neglected.

• Due to time limitations the Revit model and the Matterport scans have not been connected.

• NDT will not be used in this project.

1.10 Thesis outline

The outline of the thesis is shown below:

Chapter 1: Brief introduction to the subjects of the report along with the aim, research questions and scope.

Chapter 2: A literature review of PC bridges, thermal effects on concrete, monitoring and mate- rials, FEM, scanning and BIM.

Chapter 3: The chapter describes the method of the project. How the field work was conducted, how the data was obtained and how the FEM programs were used during the project.

Chapter 4: Results from the monitoring, FEM, scanning and BIM.

Chapter 5: Analyse of the results. The following questions were answered:

- Are the monitoring results reasonable and reliably?

- What conclusions can be made from the example data from the monitoring program?

- Are the FEM results reasonable and reliably?

(23)

- Is it possible to validate the hypothesis from the results obtained from the monitoring and the FEM?

Chapter 6: Conclusion from the project along with a discussion. Suggestions of improvements and further studies was also included in this chapter.

(24)

2 Literature review

There are some research projects dealing with the new demands of railroad traffic in Europe today. According to (Sustainablebridges, 2016) a sustainable and effective railway infrastructure is of great importance when it comes to the transportation sector in Europe. However, due to the increasing loads and speeds mentioned before, new guidelines are needed in order to tackle the new demands. Sustainable Bridges is a project that was started in 2010 with the aim to provide preparedness for the increased demands of railway bridges in 2020.

Another project with the purpose, among others, to improve the European rail system including the determination of the health of bridges is the IN2RAIL, funded by the European Commis- sion (In2rail.eu (2019)). The aim of the project is to enhance, todays and the futures, infrastructure to satisfy the 2050 estimated freight and passengers services demand growth by 80 % and 50 % respectively.

The above-mentioned projects high lights the importance of the infrastructure and particularly the rail transportation including the many rail bridges. The rising demand of today’s bridges calls for more efficient ways of determining the health, leading to prolonging the lifespan of bridges or better maintenance. The state of health is often mentioned together with monitoring.

2.1 Pre-stressed bridges 2.1.1 Principle

Hulse, et al. (2017) illustrates that the great compressive strength of the concrete can be utilized to a further extent than ordinary RC sections because of the decreased tensile stresses in the section. The compressed concrete leads to a decrease of deformation and tensile stresses in the section and the tensioning is usually done using tendons, which often are many steel wires (strands) put together. In Figure 7 different phases of pre-stressing are illustrated. The midline represents the tendon which compresses the concrete on the third beam from above and giving a positive deflection. The second beam from above illustrates the deflection of the beam when no pre-tensioning is added but a distributed load is. And the last, and fourth, beam represent the equilibrium of the prestressing and the distributed load.

Figure 7 – The different phases of pre-stressing.

(25)

If the pre-stressing force is acting at the neutral axis (where a cross-section has neither longitudinal compressive nor tensile stresses when is subjected to a bending moment) of the cross- section, it will give rise to a compressive stress contribution to the cross-section. This compressive stress contribution along with the stress distribution, that is the result from a bending moment, is illustrated in Figure 8. Notice that this stress distribution most often has an additional bending stress distribution that comes from the eccentricity of the resulting compressive force to the neutral axis of the cross-section.

Figure 8 - The stress distribution with the compressive force from the tendons acting in the centre of the section.

There are two main methods of pre-stressed concrete; pre-tensioning, meaning that the tensioning is done before the casting, and post-tensioning, meaning that the tensioning is done after the casting. According to Hulse, et al., (2017) pre-tensioning is a method were the anchorage force of the tendons are released when the concrete has reached sufficient strength. The structure then is dependent on the bond between the tendon wire and the concrete. Because of the depend- ence of the bond strength the wires of the tendons, using this method, are relatively small in diameter. The small diameter means that the tendons are more sensitive to corrosion in the concrete. In order to handle this problem factory production is recommended. Besides the pre-tensioning, there is the post-tensioning where the tendon wires are put into flexible ducts. When the concrete has reached sufficient strength the tension force is added to the tendons. Afterwards, the duct usually is filled using grouting, giving bonding between the concrete and the tendons. This option is the most commonly used method, but grouting may not be used. The bond that is created due to the grouting in the flexible ducts contributes to the ultimate strength of the section but are mostly regarded as a corrosion protection of the tendons.

Bronormer (1978) describes the four post-tension systems that was allowed in 1978; BBRV, Freyssinet, VSL and Dywidag. Because of the restriction and dimensions in correlation to the bridge dimensions there was a small number of types of these systems used. In Table 1 the systems along with their types are listed. The red lines illustrate the most used types around that time.

(26)

Table 1 - The four systems, including their types, allowed according to Bronormer (1978) (Täljsten, et al. (2019)).

System Prestressing unit Steel quality

f0,2/ft Load at failure

Dim. for anchor plate or block

Minimum distance for

anchorage¹

Minimum distance from edge to center of anchorage²

Move- ment at anchorage

Inner diameter for duct

Distance between support in

form

Concrete compressive strength at pre- stressing, minimum

[MPa] [kN] [mm] [mm] [mm] [mm] [mm] [m] [MPa]

BBRV Wires 12 6 1520/1770 600 140x140 190 (160) 100 1,0 30 < 0,8 28

Wires 22 6 1520/1770 1100 200x200 250 (220) 135 1,0 40 < 1,0 28

Wires 32 6 1520/1770 1600 220x220 305 (270) 165 1,0 50 < 1,2 28

Wires 44 6 1520/1770 2200 260x260 350 190 1,0 60 1,2-1,5 28

Freyssinet Wires 12 5 1470/1670 390 100 155 (120) 90 3,0 30 < 0,8 28

Wires 12 5 1470/1670 745 120 215 (150) 110 5,0 40 < 1,0 32

Wires 12 5 1320/1570 940 150 240 (180) 125 6,0 42 < 1,0 32

Tendon 12 12.7 1670/1860 2200 260x270 350 190 8,0 60 1,2-1,5 28

VSL Tendon 3 13 1560/1830 550 140x140 190 (160) 100 4,0 41 < 1,0 28

Tendon 7 13 1560/1830 1285 210x210 290 (250) 155 4,0 50 < 1,2 28

Tendon 12 13 1560/1830 2200 240x270 350 190 4,0 60 1,2-1,5 28

Tendon 3 ½’’Dy 1670/1860 625 140x140 190 (160) 100 4,0 41 < 1,0 28

Tendon 7 ½’’Dy 1670/1860 1460 210x210 300 (260) 160 4,0 50 < 1,2 28

Dywidag Bar 1 26 830/1030 520 130x130 180 (160) 90 0,5 32 1,8-2,5 28

Bar 1 32 830/1030 785 170 220 (210) 110 0,5 38 1,8-2,5 32

The pre-stressing may over time decrease due to creep, elastic shortening, relaxation of the pre-stressing steel and shrinkage of the concrete (Hulse, et al. (2017)). Which is problematic since the tendons carrying the pre-stressing force often have their anchorage casted into the concrete.

Otherwise, it could be a possibility to configure a new pre-stressing force in the tendons after the years of reduction. Hulse, et al. (2017) states that the use of post-tensioning of the tendons leads to a decrease of the elastic shortening of the concrete since the post-tensioning is done in steps while the pre-tensioning is made instantly in one step. It is also stated that the curvature of the flexible ducts leads to friction losses resulting in decreasing pre-stressing forces in the section.

2.1.2 Pre-stressed bridges in Sweden

Täljsten, et al. (2019) points out the historical background of the PC bridges in Sweden, stating that one of the biggest improvements to the PC bridge design code were made in the 1980’s. The improvements led to a ban on chlorides in the concrete, further developed ducts and a safer grouting process. The improvements were, however, not implied immediately why it is important to inspect these aspects in the bridges that was constructed around that time.

All Swedish bridges managed by the Swedish Transport Administaration, Trafikverket, are documented in the BaTMan (Bridge and Tunnel Management) system. The system contains not only bridge structures but piers, docks etc. from other administrations like municipalities in Swe- den. Trafikverket conducts visual inspections of all bridges every 6 year. The reports from these inspections are later added to the BaTMan system, giving a clear view on the inspection history of the bridge used to determining the maintenance strategy for each bridge for example. In the BaTMan system, 1686 PC bridges in total in Sweden are managed by Trafikverket. The age distribution of these bridges is illustrated in Figure 9. Besides these PC bridges, there are other managers of PC bridges, like the Gothenburg municipality etc.

(27)

Figure 9 - The age distribution of the PC bridges managed by Trafikverket (BaTMan).

BaTMan (2018) gives an approximation of the number of the different types of bridges since the bridges are not clearly defined in the system. It says that about 1050 are beam bridges, 350 box girder bridges, 200 slab bridges and 80 through bridges.

2.1.3 Damages to pre-stressed bridges

Shanafelt, et al. (1980) divide the damages assessment to PC bridges into three types; damage to PC strands, damage to concrete and structural integrity. Damages to the PC strands means that a loss in PC compressive force is seen and that these cause a decreasing load resistance. Damage to the concrete includes damages like cracks and spalling that could be mapped during inspection and caused by several reasons, e.g. thermal actions. The intent of structural integrity is that ana- lytic methods are used and that for example critical stresses in a section are calculated. The types of damages are also divided into three classes; minor damage, moderate damage and severe damage. Minor damage implies spalling of the concrete and cracks not wider than 3 mm not giving rise to any visible reinforcement. Moderate damage is spalling and cracks in the concrete that is wider than 3 mm and results in visible reinforcement, not including the strands. Lastly, critical damages referring to damages to concrete and reinforcement. This type could either be: cracks across the bottom flange and web, a sudden offset along of the bottom flange, loss in pre-stressing force to extent where repair is not possible, misalignment of the girder along the vertical axis, longitudinal cracks in the top flange and/or the webs reaching deep into the girder.

Bullock, et al. (2011) studied PC bridge girders constructed in Alabama where, severe damages, cracks were appearing after post-construction loads were added to the superstructure. The cracks occurred in the diaphragm walls at the continuous supports and along the bottom of the girder. Previous investigations showed that the cracking was a result of the thermal actions and not sufficient secondary reinforcement detailing. It is also stated that hairline cracks close to the supports are not uncommon in PC bridge girders.

Podolny (1985) studied a PC two cell box girder bridge with three spans of 53.6 m, 71.3 and 53.6 m with relatively short radius curvature. The cross-section varied along the longitudinal plane, the web, bottom plate and top plate thicknesses increased close to the column supports.

The PC was done using 12 tendons tensioned from both ends. A blow out of the tendons occurred when the tensioning was done in the end of the construction stage. Analysing the cause of the failure led to the conclusion that PC box girder bridges should not have too sharp curvature and the tendons should not be put too close together in the cross-section. The analysis also showed

(28)

that the concentrated force from the tendons acting in opposite direction of the arch compressive force gives rise to negative moment in the web, see Figure 10.

Figure 10 – The resulting stress distribution of the web, redrawn from Podolny (1985).

Podolny (1985) stated that, severe damage, lateral cracks may occur when tensioning is being done since the bundle of strands may flatten out causing radial forces. It is also explained that the curvature in the vertical place could give rise to longitudinal cracks in the diaphragm walls or in the webs since these often are clustered in the top of the box girder cross-section. This is causing radial forces like explained before but in another direction. In Figure 11 an example of an anchorage is illustrated showing the assumed stress distribution, the actual distribution and the damage caused by the wrong assumption.

Figure 11 - The design assumptions correlations to the reality, redrawn from Podolny, (1985).

Damages of PC bridges caused by thermal actions

Imbsen (1985) studied the Jagst bridge in Germany, which was a PC box girder bridge that spanned totally 52.2 m in two spans, see Figure 12. Around five years after the construction of the bridge was finished, severe damages, cracks were found in one of the webs along the longitudinal axis of the bridge. The cracks had a width of around 5 mm. The reason of the cracking was at first addressed as temperature induced. Later, it was discovered that large pre-stressing forces and not sufficient amount of second reinforcement in the crack areas were contributing to the thermal effects causing the cracking of the concrete. The large pre-stressing forces came from the fact that big ducts had been used, leading to concentrated tensile stresses in the webs. The inspection team concluded that the stirrups was not enough to handle these tensile stresses. As a solution to this it was suggested that thick webs in PC box girders should be avoided and that the second reinforcement detailing should be increased.

(29)

Figure 12 – The cross-section of the Jagst Bridge in Germany, redrawn from Imbsen, (1985).

In Figure 13 common temperature cracks of PC box girder bridges in 1981 are visualized by inspection of several bridges Zichner (1981). These cracks were found in the same places in the webs and the bottom plate of the box girder bridges, in the transversal direction. Most of the bridges were curved and contineous multi-span bridges and the crack width were most often around 0.2 mm.

Figure 13 - Typical thermo cracks in the 1990's in PC box girder bridges, redrawn from Zichner (1981).

Around 1985 Imbsen (1985) studied four PC box girder bridges in Colorado, United States, that had the same problems with cracks occurring in the webs and bottom plate of the box girders.

The longest two bridges were approximately 228 m long and the other were approximately 158 m and 137 m long in total. The shortest bridge had three continuous spans and had the greatest amounts of cracks, maximum crack width was around 3 mm. During an inspection of the bridge the temperature were measured at three points in the box girder bridge. After analysing the temperature along with the crack widths over time, it was concluded that the structural performance was impacted by the temperature distribution in the box girder. In Figure 14 the shortest bridge is illustrated along with the damages. The other three bridges that were studied had similar cracks.

(30)

Figure 14 – The shortest of the studied Colorado PC box girder bridges, redrawn from Zichner (1981).

White, et al. (1980) examined the inspections of totally 64 concrete box girders with a length greater than approximately 107 m without joints for expansion. The aim of the study was to determine the shrinkage of the concrete among others, but cracks were discovered during the process. Cracks that were not supposed to occur. These cracks were located in 20 of the curves, horizontally over the supports and in the tension zone as the bridge is subjected to thermal shortening. Small cracks were also found in the abutments. The cause of these cracks was hard to determine but it was concluded that the temperature effects must have contribution to the cracking of the concrete.

Another common problem to PC box girder bridges was described in Roberts-Wollman, et al.

(2002), tensile cracking in the deck of the bridge due to high tensile stresses induced by thermo loads. The transversal distribution of the temperature in the cross-section was not regarded by the code at that time was concluded.

2.1.4 Abiskojokk bridge

The Abiskojokk bridge is a post-tensioned box girder bridge located in the north part of Sweden along the Iron Ore Line that stretches from Luleå – Kiruna – Abisko – Narvik (Norway). It was constructed in 1978 and carries a single rail track. The superstructure carries the greatest load when the iron ore trains are passing and since the construction of the bridge the axle load has been increased from 140 to 325 kN in different stages (Coric, et al. (2018)). The bridge is illustrated in Figure 15.

(31)

Figure 15 – The plane and sectional view of the bridge, part of drawing from BatMan (2018).

The width of the concrete deck is 6900 mm, the height of the box girder is 2100 mm and the width of the box girder is 3500 mm. At the column supports the cross-sections are thicker than the rest of the box girder. There are also a beam creating a diaphragm wall at the column supports.

At these diaphragms, the walls are 450 mm thicker on each side and the floor is 200 mm thicker.

Figure 16 – The cross-section of the box girder, part of drawing from BatMan (2018).

The bridge is prestressed using 12 tendons of the type BBRV. Each wire consists of 44 small wires (strands) with a diameter of 6 mm. The 0.2 tensile stress limit is 1520 MPa and the failure strain is 35 %. Surrounding the tendons are ducts with the diameters of 60/67 mm. Along with the tendons two types of anchorages were used, 4S-v and 4S-y according to the former Swedish

(32)

Transportation Administration, Statens Vägverk. In Figure 17 the first anchorage type (used in the midspans) is illustrated along with the cross-section of the tendon.

Figure 17 - The anchorage 4S-y along with the section of the tendon, part of drawing from BatMan (2018).

The bridge superstructure and piers were casted using K40 which is an old nomination of concrete quality and the reinforcement bars that was used was of the quality Ks 40, Ks 40s and Ks 60, which also is a former nomination of the steel reinforcement. Strengthening of the foundations to the columns were made during 2001 where an approximately 950 mm thick layer of reinforced concrete where added to the existing foundations.

Figure 18 - The crack groups location inside the box girder, redrawn from Inspektionsrap- port (2016).

The latest inspection (särskild), Inspektionsrapport (2016), mapped cracks in the box girder.

The crack orientation and information can be found in Figure 18 and Table 2.

(33)

Table 2 - Cracks discovered during Särskild inspection 18th of August 2016, modified from Inspektionsrapport (2016).

Measuring point

Crack width

(mm) Location Number Remark

1 0.1

Right box girder wall, second support (in-

side) 1

2 0.4 Column 2, right side 1

3 0.25

East side of column, appr. 5.6 m from

ground 2

101 0.1 Right box girder wall, third span (inside) 6 102 0.1 Right box girder wall, third span (inside) 5

103 0.3 Right box girder wall, second span (inside) 1 3.5 m long 104 <0.1 Right box girder wall, second span (inside) 18

105 0.15 Right box girder wall, second span (inside) 17 106 0.15 Right box girder wall, second span (inside) 12 110 <0.1 Left box girder wall, third span (inside) 8 111 0.15 Left box girder wall, third span (inside) 6 112 0.1 Left box girder wall, second span (inside) 15

113 0.3 Left box girder wall, second span (inside) 1 2 m long 114 0.1 Left box girder wall, second span (inside) 13

115 0.1 Left box girder wall, first span (inside) 15

201 0.2 Man holes above column supports 8

301 0.1 Box girder roof, second span 1 3 m long

302 0.15 Box girder roof, first span 3

19 m long in total

401 0.25 Foundation column support 1 1 4 m long

When the current axle load, 325 kN, was to be implemented a classification of the bridge was conducted. The classification done in 2016, Bennitz (2016), showed that the bridge was able to handle the new axle load.

2.2 Thermal effects

Thermal effects considered in structural design normally consists of three stages; when the casting is taking place the hydration process results in heat rising, the cooling of the concrete after the hydration process meaning that the temperature in the structure correlates to the ambient temperature and, lastly, the climate giving rise to temperature variations during the service life of the bridge. Temperature variation ambient to superstructures like bridges will cause contraction or expansion of the bridge, depending on if the temperature is getting higher or lower. If a sub- structure is restrained, which it often is, stresses caused by the temperature variations will occur.

The stresses derived from the third stage of the thermal effects was the focus in this project.

Rajeev, et al. (2016) reported that roof tiles in Australia has been exposed to cracks in the concrete due to the thermal effects from the climate. One of the dangers with cracks in concrete is that the concrete protection of the reinforcement may disappear, meaning that the reinforcement will corrode faster and in the end the structure will lose load capacity, or even collapse. It was also stated that the most common cracks in concrete are caused by shrinkage and temperature gradients. The thermal cracks are mostly caused by temperature gradients created by the solar radiation and in an early stage of the construction because of the hydration process taking place as the concrete is hardening. The solar radiation may cause large temperature gradients because of the relatively low heat conduction of concrete. When one side of an element is exposed to the

(34)

radiation and the other side is not a temperature gradient is hence created. The temperature gradient in the material gives rise to internal stresses that causes for example tensile stresses along an edge giving rise to cracking of the concrete, because of the low tensile strength of the concrete.

Tang, et al. (2012) stated that another common cause of the thermal cracking of the concrete is the different parts of the concrete not matching in terms of heat conduction. Meaning that the different parts of the concrete structure will have different behaviours to changes in the temperature and stresses induced by different heat conduction properties will occur.

In Imbsen (1985) it stated that the temperature stresses may be greater than the live loads applied to the bridge and that codes in the 1990’s and back did not cover the problems of thermal loads to concrete bridges. It was also found that the solar radiation can result in non-uniform temperature gradients going through both the width (transversal) and the depth (longitudinal) of the cross-section. It is explained that the nonlinear temperature distribution can lead to cracking of the concrete in two ways:

• Induced by the temperature, the distortions of the concrete giving rise to an internal bending moment around the longitudinal axis.

• The predisposition of plane sections remaining plane gives rise to a prohibition of the non -linear distortions, resulting in stresses close to the supports.

It was noted that the general code in the United States (US), AASHTO, at that time mostly regarded the longitudinal expansion/contractions while the transversal was left not stated.

Imbsen (1985) denoted that the thermal effects on a bridge contains a combination of factors like the solar radiation intensity, the depth of the structure, the wind power, the humidity levels ambient to the superstructure and so on. These factors are provided in Figure 19.

Figure 19 - The factors affecting the temperature gradient of box girder bridges.

Roberts-Wollman, et al. (2002) described the mechanism behind the common tensile cracks in the top part of the deck of PC box girder bridges. They were often caused by a negative temperature gradient applied to the section. An explanation is provided in Figure 20 and Figure 21, where for example a negative temperature gradient applied to the concrete deck, which gives rise to a corresponding strain distribution. The average shortening, the curvature and the temperature gradient corresponding strain distribution is summarized and multiplied with the Young’s modulus of the box girder section, resulting in a stress distribution of the section, seen in Figure 21.

Notice that the top deck and the bottom flange of the box girder cross-section will be in tension.

Monitoring and Modelling of the Abiskojåkka Bridge

Monitoring and Modelling of Abiskojåkka Bridge

Kasper Furenstam

Kasper Furenstam

MONITORING AND MODELLING OF THE ABISKJÅKKA BRIDGE

Master thesis Väg- och vattenbyggnad, inriktning Konstruktion, 30,0 hp, X7010B

Abstract

Sammanfattning

Preface

List of Abbreviations

List of Notations

Contents

1 Introduction

BIM

2 Literature review