Holistic approach in engineering design - controlling risks from accidental hazards in bridge design

Sammanfattning: Popular Abstract in English A central requirement in the design of bridge, or any structure, is that they are safe and do not fail unexpectedly. To ensure that such occurrences are acceptably rare involves predicting the various ways in which the bridge can fail and ensuring that provisions are made to prevent or control these potential failure. In other words, the risks associated with failure should be controlled such that they are within acceptable limits. The broader issue investigated in this thesis relates to the manner in which risks of failure are treated in engineering design of structures – specifically bridges – to ensure their safety. Historically, the safety of a structure was assured by relying on past successful practices and by using factors of safety to account for any uncertainties. These factors alter the design values of loading or structural capacity to ensure a sufficient margin for error against the unknown. Historically, the selection of such factors was based on experience and engineering judgment and failures from less well understood phenomena had higher factors of safety. However, in the middle of the 20th century rational approaches to the selection of safety factors and of the treatment of structural safety were developed which eventually started replacing these previous methods. Currently, design codes are the main instruments used in regulating procedures for verifying that a structures design is adequate in terms of safety against collapse. This includes specifying the magnitude of safety factors to be used during design. Such codes are written by code writing bodies and help provide assurances to society and other stakeholders that a design is safe and compliant with building regulations. There are, however, some important limitations of design based on code-compliance. These limitations are such that revisions of codes do not suffice in remedying the situation; a complementary approach is needed. The main outcome of this thesis is a framework for an approach which is intended for treating risks inadequately covered by the design code. This approach focuses on the design of bridge structures to withstand accidental or extreme loading situations – or hazards. These situations are rare and their magnitude can be very significant compared to more conventional loading situations such as the weight of traffic on a bridge. The approach also focuses on the how to consider the risks from these hazards during the early conceptual stages of design – in which case there are usually multiple competing technical solutions. The main advantages with the approach are: 1. The approach is holistic and broadens the scope of assessment – while codes are focused on the safety of single structural elements 2. The approach can be applied during the conceptual design stage – while the codes are only applicable during the detailed design stage 3. The approach complements codified design instead of competing with it In developing the approach further, three crucial aspects were identified and investigated. To start, the usefulness of the approach during the conceptual design phase was tested by applying it for an actual construction project in Sweden. Second, the importance of broadening the scope was highlighted when conducting a more detailed assessment of a specific hazard – train collisions – to a second bridge case. Finally, risk modeling approaches were investigated in detail for the case of heavy vehicle collisions to bridge supporting structures. It was found that current codified approaches relating to this hazard are inadequate and an alternative design method was developed using stochastic modeling approaches – where uncertainties related load and structural behavior are directly considered. Furthermore, a simplified method for assessing the risks related to this hazard was developed.