Quay walls are an essential part of port infrastructure, providing diverse functionalities for the quay owner and users. In order to guarantee the structural safety of existing quay walls, owners might want to maintain their quay walls or decide to reassess their safety.

For various reasons, quay wall owners want to lower maintenance costs and postpone investments on their quay walls. Current methods for the reassessment of quay walls are partly depending on stochastic variables, which might lead to a rejection of the quay wall, while it is still safe to use them. Therefore, this research aims to develop an assessment method for existing quay walls based on observations of their real behaviour. This leads to the following research question:

How can existing quay walls in seaports be assessed using real-time data of their deformation behaviour?

In order to answer this question, a method was developed in this research for the assessment of quay walls based on their real behaviour. This method consists of 6 steps: .
1. The first step is the preparation of the data. All data is resampled to hourly values, in order to create a dataset with equal timestamps;
2. In the second step, Bayesian regression is used to create a prediction model of single measurement points on the quay wall;
3. The third step is to prepare the prediction models of the measurement points for the assessments;
4. The fourth step is the short term assessment of the quay wall. Real-time monitoring data is compared to the prediction model to assess if the quay wall shows safe deformation behaviour;
5. The fifth step is the assessment of the difference of the prediction models made for different measurement points on a single quay wall;
6. The sixth step is the assessment of the remaining capacity and the remaining lifetime of a quay wall.

To showcase the method and to verify and validate various aspects of the method, an example case was used. This case consists of four adjacent quay wall sections, that have been rejected by means of load/resistance based calculations. The developed assessment method was used to reassess these quay walls.

The results of this research show that the elastic behaviour of a quay wall is mainly influenced by the air and water temperature, the water level in the port, the groundwater level and direct loading. A prediction model created with Bayesian regression based on these causes was able to predict the quay wall behaviour accurate enough to use it for the assessments of the last three steps of the assessment method.

The application of the assessment method to the quay walls of the example case show that three of the four quay wall sections have enough resistance. One quay wall section shows questionable behaviour, which can be related to an early failure during the construction works.

The differences between the models of the different measurement points on the quay walls of the example case are all explainable by their location on the quay wall and the history of the quay wall. The most important explainer for the difference of the models of the different measurement points was the quay wall section on which the measurement points were located.

During the monitoring project, reinforcing measures were taken to try to reduce the ongoing quay wall deformation. The results of this research show that the measures were successful for most quay wall sections. Only at the section that was already showing questionable behaviour the situation worsened after the application of the reinforcing measures.

The conclusion of this research is that the six-step method explained above can be used to assess the deformation of a quay wall. Recommendations for future research are the broader application of the assessment method, the application of the assessment method with other monitoring techniques and the disentangling of the linear plastic trend in its deterministic causes.