Research Fellow at AMS Institute, and PhD candidate at TU Delft, Mart-Jan Hemel is investigating the condition of these centuries old quay walls by doing exactly what the local authorities are trying to prevent: put pressure on them until they collapse.

At first glance the former Zuidergasfabriek terrain in Amsterdam looks a bit like a construction site. And yet from the end of February until mid April the old dock on the Amstel will be the scene of a very important experiment.

In this (Dutch) video Jelle van Ophuizen (Iv-Infra), Erik Hutcheson (Gemeente Amsterdam), and Mart-Jan Hemel (TU Delft and AMS Institute) more about this innovative test load. This video was commissioned by Iv-Infra in cooperation with Gemeente Amsterdam, AMS Institute and TU Delft.

“‘We will be able to determine here how much pressure quay walls can take before they bulge or collapse. It is, of course, impossible to conduct tests like this in the city centre. That is why the council has found us a piece of derelict land to use as an open air lab. In a few years’ time there will be office blocks here, and apartments.’”

Mart-Jan Hemel

Former Research Fellow at AMS Institute

Photo by Frank Auperle

Vulnerable quay walls

It is no coincidence that the experiment is taking place in Amsterdam. The capital has a whopping 200 kilometres of quayside, all built on wooden piles. Due to their great age, proper documentation is often missing. It is not always easy to pinpoint the weak spots , or determine how long a construction will last without intervention. ‘In 2020 the quay wall of the Grimburgwal, which is privately owned, collapsed into the canal over a length of 30 metres, creating a potentially dangerous situation. Amsterdam city council has an ongoing programme of strengthening quay walls and shoring up potential weak spots but wants to prioritise intervention based on more precise information.’

Predicting quay wall strength

In order to predict how damage to the quay wall develops over time more data are needed, Hemel says. ‘Hard figures will tell you if a quayside is due for a repair job. Subsidence risks cannot always be spotted with the naked eye, but, at the same time, bulging or cracking does not always mean the quay wall is on the point of collapse.’
The experiment on the Amstel will give Hemel the data he needs to make better predictions. The civil and hydraulic engineering engineer developed a model which provides an insight in the stability of quay walls.

“It consists of mathematical equations and physics formulas. By feeding data into the model – such a soil type, the number of wooden supports the wall is resting on or the pressures acting on it caused by traffic or concrete blocks – you can calculate how long a quayside will remain stable.”

Mart-Jan Hemel

Former Research Fellow at AMS Institute

Validation test

But before it can be used the model needs to be calibrated and validated, Hemel says. ‘That is why we need to do the experiment. The results will tell us if my theory-based suppositions are compatible with what happens in practice. It may be that my model predicts an expansion of 5 centimetres under a given pressure but that in practice the wall expands by as much as ten centimeters. If that is the case I want to know what is causing the difference. It may be that the input is incorrect, or that I have to add certain processes and parameters to the model.’

Quayside divided into four test sites

The experiment took 2.5 years to prepare (see box). Hemel: 'In the end we decided to divide the quayside into four quay wall testing sites and one pile testing site. This will enable us to repeat the experiment under different circumstances. Water depth, the location of the load, the composition of the soil behind the quayside or the presence of tree roots in the underlying layers are all variables that play a part. In some of the sites we remove the wooden piles to weaken the quay wall. To prevent the whole thing from collapsing the individual testing sites have been separated by a distance of a couple of metres. These spaces have been shored up by a construction of pipes and sheet pile walls.’

Photo by Frank Auperle

Sonar equipment and underwater photography

Before load testing can begin we need to know what the quayside looks like below the waterline, Hemel says. ‘The condition of the quayside is determined by the number of piles, their thickness and how deep they are positioned. To find out divers have used sonar equipment and underwater cameras. Monitoring systems have been fitted on and underneath the quay wall so we can see what happens to the elements that make up the construction during the load testing.’


The first container was placed at the first test site in February. Despite a long and meticulous preparation Hemel was a little nervous. ‘This research project offers opportunities but there are risks as well. What we did not want was to damage  infrastructure such as water pipes in case of a collapse. So safety was paramount.’

First phase

To increase the pressure a water bag inside the container was filled bit by bit. Hemel: ‘This is done in steps of 30 centimetres, until the quay wall stabilises. We can see the deflection of the quay wall and the subsidence of the surface level on the monitoring screens.’ The whole process is a painstaking one, Hemel says. ‘The bigger the load, the longer it takes for the wall to stabilise.’

“We found some ‘interesting deviations’ from the model predictions following the first test but it’s too early to draw any hard and fast conclusions. So far, the execution of the experiment is going just the way we envisaged. Now we will turn our attention to the other test sites. We may weaken the quay wall a bit, or increase the load a bit faster. We can only hope for the moment when it collapses. We’ll see if and when that happens.”

Mart-Jan Hemel

Former Research Fellow at AMS Institute

Mart-Jan Hemel is a PhD candidate at TU Delft and Research fellow at AMS Institute.

Source: This 'Story of Science' (also available in Dutch) is derived from the TU Delft Stories page. Read more TU Delft stories here >>