The need of exploring efficient solutions to today’s engineering problems is becoming essential in the current market development. Soil Steel composite bridges (SSCB) are considered well competitive in terms of their feasibility and constructability. The primary objective of this study is to provide a comprehensive comparison study for two known design methods of SSCB, which are the Pettersson-Sundquist design method (developed in Sweden) and the Klöppel & Glock design method (developed in Germany). Moreover, in the goal of having better behaviour understanding for SSCBs, the study also include finite element modelling (FEM) using PLAXIS 2D of three case studies and compare model results with field measurements.   The design comparison deals with the design concepts, assumptions and limitations for both design methods, where full design procedures are implemented and compared for a defined case study.   The results of the FEM analysis show rational outcome to the field measurements for structural response during backfilling and close results for ordinary loading as well. While the design comparison shows how the different approach in both design methods in limitations and design assumptions has important impact on the results, where soil failure in the Klöppel & Glock design method can be controlling the design for low heights of cover, whereas formation of crown plastic hinge is more controlling in the Pettersson-Sundquist design method. However, and in general, the Pettersson-Sundquist design method require more steel in low heights of cover while it is less demanding for higher soil covers compared to the Klöppel & Glock design method.

Nowadays, many different efficient solutions are being studied to solve engineering problems. Inside this group of solutions we can find the Soil Steel Composite Bridges (SSCB) as an alternative to traditional bridges. SSCB are being used more often every day and they are showing themselves as competitive structures in terms of feasibility and constructability. This project was started to achieve two different goals. The first one was to create a general database of SSCB including few selected tests all around the world and the second one was to compare and discuss full scale tests using the Pettersson-Sundquist design method. To create the database and the following comparisons, twenty-five different full scale tests were used. From this tests all the necessary information was extracted and used to create the database. After creating the database, the project continued with the discussion and comparison of the full scale tests. Specifically those discussions and comparisons were related to the resistance of the soil (the soil modulus) used in the construction of the SSCB. All the values of the different soil modulus of each full scale test used in the comparisons were calculated using the Swedish Design Manual (SDM). Two different types of soil modulus were calculated in this project using SDM, ones are the soil modulus back calculated using the values reported from the live load tests performed on the culverts and the others are theoretical soil modulus calculated using the detailed information of the soil. The report continues with the explanation of the different conclusions ended up with during this project. It can be highlighted within this group of conclusions, the one related to the importance of reporting all the necessary information from the full scale tests including the soil parameters, the measures of the culvert, the cross sectional parameters and the vehicle dimensions among others. Another important conclusions are the effect of using the slabs over the top of the culvert and how it would effect to the sectional forces over the culvert and also the limitations using method B of the SDM regarding the type of soil used as backfilling Finally, the project finishes explaining some proposals for future research about other fields of the study of SSCB.

Soil-steel composite bridges are for many reasons very favorable bridges to build. Up to a certain span length they are economical, practical to transport and simple and quick to build. Especially in remote places, with unfavorable sloping terrain, this can be of great interest when the transport of material can be difficult and costly. The swedish manual for design of soil steel composite bridges was presented by Lars Pettersson and Håkan Sundquist in the year 2000. It is today used in Sweden and other neighboring countries as the main manual for the design of soil-steel composite bridges. The design manual is however only valid for longitudinal slopes up to 10%, which is low in comparison to natural sloping hillsides. The purpose of this thesis is to study the structural behavior of soil steel composite bridges in sloping terrain, with the use of the finite element software PLAXIS 2D. Two case studies of one low arch culvert bridge and one pipe arch culvert are studied and later modeled in PLAXIS. The two case studies does not have sloping terrain, but are used to assure that the FEM-models behaves correctly during the backfilling process in PLAXIS. The analysis studies the change in sectional forces in the bridges during an increase inslope above the structures. The values from the FEM-models are compared to field measurements from the case studies, as well as to values calculated in accordance to the swedish design manual. This is done to see how well the different approaches compare to each other, and, since the design manual does not consider slopes exceeding 10%, it is interesting to investigate if the SDM-calculated values still appear valid for slopes larger than 10%. In addition to sectional forces, the slope stability of the FEM-models is also evaluated and compared to analytical values calculated using the ordinary method of slices. This slope stability study focuses on if the bridges affect the slope stability safety factor. The thesis also investigates if these bridges are more sensitive to increasing slopes from a slope stability or sectional forces point of view. I.e if the slope inclination magnitude will be determined based on the sectional forces in the structure or the slope stability of the soil surrounding the structure. The criteria that are researched to come to this conclusion are for example the slope stability safety factor, or the possible yielding of the culvert wall due to too large sectional forces in the structure. The results show that for low cover depths, the slope stability safety factor decreases slightly when a structure is introduced to the slope. For larger cover depths, read 3 meters or more, the slope stability seems unaffectedby the structure. The results also shows that the sectional forces in the structures compares well to both measured and calculated values for slope inclinations up to 30% for rather small cover depths, i.e 1 meter. For slopes larger than 30% the sectional forces grows and no longer reflect the measured and calculated values. However, since only two case studies are performed, these conclusions might not be valid for different profile shapes. Additionally there are indications that the low arch bridge is more sensitive to slopes than the pipearch culvert. / Rörbroar, eller “soil-steel composite bridges”, är av många anledningar fördelaktiga att bygga. Upp till en viss spännvidd är de ekonomiska, praktiska att transportera samt enkla och snabba att konstruera. Särskilt vid vissa svåråtkomliga platser, med ofördelaktigt sluttande terräng, kan det vara av stort intresse när materialtransport annars kan vara både komplicerat och kostsamt. Den svenska manualen för dimensionering av rörbroar presenterades av Lars Pettersson och Håkan Sundquiståret 2000. Manualen används idag i Sverige samt angränsande länder som den huvudsakliga dimensioneringsmanualen för rörbroar. Dock är den svenska manualen enbart giltig för lutningar upp till 10%, vilket ären låg lutning i jämförese med många naturliga sluttningar. Syftet för detta arbete är att studera det statiska beteendet hos kringfyllda rörbroar i lutande terräng,med hjälp av det finita element programmet PLAXIS 2D. Två fallstudier av dels en låg bågbro, dels en sluten kulvertbro, studeras och modelleras senare i PLAXIS. De två fallstudierna har inte sluttande terräng, men används för att säkerställa att FEM-modellerna beter sig korrekt under återfyllnadsfasen I PLAXIS. Analysen studerar förändringen av tvärsnittskrafter i broarna när sluttningen de belastas av ökar. Värdena från FEM-modellerna jämförs mot mätdata från fallstudierna, samt mot värden beräknade enligt den svenska manualen. Detta utförs för att se hur väl de olika metoderna överensstämmer med varandra, och, eftersom dimensioneringsmanualen inte tar hänsyn till lutningar överstigande 10%, är det intressant att se om deSDM-beräknade värdena fortfarande gäller för lutningar större än 10%. Utöver tvärsnittskrafter utvärderas släntstabiliteten hos FEM-modellerna, och dessa resultat jämförs sen mot analytiska värden beräknade enligt lamellmetoden, eller “the ordinary method of slices”. Släntstabilitetsstudien fokuserar på hur broarna påverkar säkerhetsfaktorn för släntstabilitet. Avhandlingen undersöker även om dessa broar är mer känsliga för ökande lutningar från en släntstabilitets- eller tvärsnittskrafts-synvinkel, dvs. om lutningens största storlek kommer att avgöras baserat på de tvärsnittskrafter som uppstår i konstruktionen eller på släntstabiliteten i marken runt konstruktionen. För att bilda oss en uppfattning om detta jämförs exempelvis säkerhetsfaktorn för släntabilitet mot ett eventuellt brott i stålprofilen orsakat av för stora tvärsnittskrafter i konstruktionen. Resultaten visar att säkerhetsfaktorn för släntstabilitet, vid liten överfyllnadshöjd över hjässan, minskar något när en bro byggs i en sluttning. För större överfyllnadshöjder, läs 3 meter och högre, verkar släntstabiliteten vara oförändrad vid introducering av en bro. Resultaten visar även att tvärsnittskrafterna i konstruktionerna överensstämmer bra både mot mätdata samt beräknade värden, för lutningar upp till 30% och små överfyllnadshöjder på exampelvis 1 meter. För lutningar större än 30% växer tvärsnittskrafterna och är inte längre jämförbara mot varken mätdata eller beräknade värden. Dock bör det tilläggas att eftersom endast två fall har studerats utförs så behöver dessa slutsatser inte gälla för andra profiltyper. Utöver detta finns indikationer på att den låga rörbron är mer känslig för en ökning i släntlutning i jämförelse mot denandra bron när det gäller tvärsnittskrafter.