NEW HOSPITAL in the EXPO 2015 AREA (Milan, Italy)

The wind tunnel campaign was developed to characterise wind-induced pressures on the building façades, with specific focus on the cantilevered façade appendages—architectural “fins” protruding from the elevation—which can become strongly load-bearing or uplifting under near-grazing winds. Alongside this primary objective, the pressure-tap layout was expanded to cover additional façade and roof portions and locally densified around balcony walkways and recessed zones, in order to capture local effects and provide design-ready information for specific façade sub-areas and details.

The tests were based on time-resolved pressure measurements, carried out under a reproduced atmospheric boundary layer. The incoming flow was characterised through dedicated profiling of mean wind speed and turbulence intensity, using an automated traversing system. Because the models had to be built at sufficiently large geometric scale to resolve façade details, the longitudinal integral length scale could not be perfectly matched over the full height of the building; this limitation was considered of minor influence for the intended deliverables, which were primarily local pressure fields and local peak actions on façade elements rather than global dynamic response. The overall approach and the measurement strategy were therefore tailored to maximise reliability and accuracy of local pressure coefficients under multiple wind directions.

The experimental program was organised in two complementary phases, using two rigid physical models that shared the same aerodynamic intent but served different roles within the campaign. A first, large-scale partial façade model was employed for exploratory testing, while a second model of the full building—including surrounding context—was used for final design conditions.

The exploratory Model A reproduced a representative façade portion at larger geometric scale, enabling rapid iteration on pressure-tap placement and on the measurement strategy required to reconstruct the pressure field over the cantilevered appendages. This model was conceived specifically to understand the pressure mechanisms on the protruding elements, including the pressure difference between their exposed and sheltered faces, and to identify where a finer spatial resolution was necessary to capture peak loads. The model also allowed controlled variations of local boundary conditions around the façade portion, including lateral confinement options intended to modify the local three-dimensionality of the flow and support a conservative interpretation of façade pressures. In addition, two limiting configurations of the horizontal supporting surfaces were investigated to qualitatively assess the sensitivity of the façade appendages to permeability effects. The pressure taps were arranged along multiple vertical alignments to capture pressures on walls, parapets, and on both sides of the façade fins. The exploratory phase was supported by qualitative flow visualisation (e.g., tufting) to help interpret recirculation zones, reattachment regions, and the flow paths responsible for local suctions and over-pressures.

The definitive Model B reproduced the entire building at a smaller scale suitable for representing both the architecture and the relevant surrounding urban context. This model provided the pressure coefficient dataset used for design. The building was installed on a rotating turntable to measure wind loads as a function of wind direction, and it was tested within an extended model of the surrounding built environment, defined over a sufficiently wide area to include aerodynamic interference effects from nearby obstacles. Two alternative “surrounds” were considered: one representative of the current context and one consistent with a plausible future development scenario. This allowed the campaign to quantify how shielding, channeling, and wake effects from upwind buildings could modify the façade loading pattern, often reducing mean pressures in sheltered zones while increasing peak excursions due to unsteady wake impingement.

Given the high number of pressure points required to describe both the façade appendages and selected façade/roof regions, the measurements were performed in multiple acquisition sets, each capturing a subset of taps simultaneously. This approach preserved adequate sampling performance while maintaining the spatial detail needed for local design, and it was complemented by dedicated high-resolution acquisitions targeted specifically at the cantilevered elements. The tap distribution on the full model followed the outcomes of the exploratory phase, with denser instrumentation where pressure gradients and peaks were expected, such as around corners, parapets, balcony recesses, and the cantilevered fins.

For the cantilevered elements and selected glazed parapets, both external and internal pressures were measured to directly obtain net pressure coefficients, ensuring that the design actions reflect the true pressure difference across the components. The pressure signals were processed to provide mean values as well as design-relevant peaks. Because pressure taps provide point measurements and may not permit true spatial averaging over the tributary area of secondary components, peak evaluation was performed through time-window averaging as a practical surrogate to represent the loading relevant to components of finite size, while retaining the contribution of turbulence and wake unsteadiness. The final deliverables were provided in a format suitable for design workflows, including pressure coefficient maps and digital datasets, enabling the derivation of load envelopes for the façade appendages and for local façade and cladding zones.

Overall, the campaign showed that the surrounding built environment can meaningfully influence façade loading, particularly for near-grazing winds along the elevations. The two-model strategy proved essential: the partial façade model enabled a physically informed and efficient definition of the pressure measurement layout on the most sensitive façade details, while the full-building model, tested with different environmental scenarios and wind directions, provided the comprehensive dataset required for final design. As with all wind tunnel studies, the results are valid for the tested geometry and the investigated surroundings; significant architectural changes or major alterations in the nearby context require dedicated reassessment.