[1][2] He contributed to several major water engineering projects in the Netherlands, notably the drainage of Walcheren at the end of the Second World War, dike restorations following the 1953 North Sea flood, and the Deltaplan.

A proponent of integrating statistical methods into engineering, van de Velde advocated for probabilistic approaches to assess safety and manage uncertainty in the design of flood defences.

In his 1980 farewell lecture at Delft University of Technology, he emphasised the limitations of deterministic safety factors and underscored the importance of using probabilistic techniques, such as Monte Carlo simulations, to model risks and failure probabilities in complex systems.

[7][8] Van de Velde was appointed as lead engineer for the Schelphoek breach on 27 April 1953, which had become so deep due to strong ebb and flood currents that it could not be closed using traditional methods.

Initial studies examined the hydrodynamic and geological conditions of the site, with extensive tidal calculations and laboratory experiments conducted to assess the forces acting on the breach and predict the behaviour of water flow during the closure process.

[6] Van de Velde also played a significant role in the closure of the breach at Ouwerkerk, which involved the use of 11,500 workers, 4 Phoenix caissons, as well as a number of tugboats and stone dumping vessels.

However, following public protests from environmental activists and fishing communities, the Den Uyl cabinet decided in 1974 to make major alterations to the project, thereby requiring a partially open design.

Van de Velde emphasised the importance of probabilistic methods, such as the Monte Carlo approach, to account for uncertainties in factors like material strength, environmental loads, and design parameters.

Drawing from the history of Dutch flood defences, he highlighted key advances in dike safety after the 1953 disaster, including the integration of statistical insights into design standards.