The Full Probabilistic Design Method of the Storm Surge Barrier Near the Port of Rotterdam, The Netherlands

After the storm surge disaster in 1953, which caused more than 1800 casualties in the Southwestern part of The Netherlands, a large dyke-strengthening and coastline-shortening programme was agreed upon and laid down by law. Work on the first projects commenced in the early sixties of the last century and the last phase of the programme was planned to start in 1990 and comprised of the dyke-strengthening programme in the Rhine Delta upstream from Rotterdam. This large project encountered growing public resistance as the required safety standards were established at the expense of both social and cultural values as well as ecological values. A feasibility study was started to ensure the required safety requirements of a storm surge barrier. The outcome was positive and the project was started in 1990 and was completed in 1996. In 1987, six contractors were invited to tender for the design and construction of a storm surge barrier, with only four “demand” specifications: (1) Reduction of the design water level in Rotterdam by 1.6 metres. (2) Reduction of the design water level 25 km (15 miles) upstream by 0.6 meter. (3) Lifetime of 100 years. (4) No obstacles to navigation. This set of requirements pertained to failure criteria. Based on this set of requirements, a full probabilistic method was adopted for the design of the storm surge barrier. A breakdown was made, starting from the basic probabilities of failure. The breakdown was based on failure trees with parallel and serial connected components and elements. In that way the design engineers were provided with centrally distributed failure criteria. This full probabilistic method, however, did not appeared to be adequate for several reasons. After a few months the full probabilistic design method was changed into a semi probabilistic method. Nevertheless, for the assessment of the load cases, a probabilistic approach was used, but for the design work on components and elements a traditional method introducing partial safety factors was used. Throughout the design period it was very difficult to prove that the actual designed system as well as, the designed sub systems and designed components met with the basic failure requirements. In order to avoid discussions, the designers embraced higher limits for their dimensioning calculations, resulting in a safer and more reliable storm surge barrier than was initially required.