Critical measurement issues in the use of wire potentiometers for the structural health monitoring of slender structures: the case of the Duomo di Milano main spire

During the restoration of the Duomo di Milano main spire, a monitoring system has been set up in order to have an early detection of potential risks of the spire. A scaffolding for the restoration activities was mounted around the spire, and it had a section exposed to the wind actions much wider than the spire itself. Any possible contact between the spire and the scaffolding had to be prevented, even under the strongest winds expected in Milan, since the stability of the extremely slender marble spire could be compromised in case of contact. A safety gap between the two substructures was granted during the construction phase; however, the deflection of the spire could not be reliably predicted by numerical models. Therefore, it was decided to continuously monitor the actual gap value, together with other measurements, helping in a thorough understanding of the spire and scaffolding movements. The gap monitoring was measured by three-wire potentiometers, mounted to measure the relative position between the scaffolding and the spire in the section where the gap was the narrowest: the top balcony. This measurement was considered one of the most critical for the safety of the spire; therefore, every choice has been aimed at redundancy and safety. Moreover, any systematic effect has to be compensated for. A main issue related to the use of wire potentiometers is that these sensors estimate the amplitude of the displacement but does not provide any information on its direction. Since the horizontal displacement of slender structures can be in any direction, according to the wind and sun radiation conditions, the output of a wire potentiometer is also affected by displacements in the direction normal to its sensitivity axis, leading to a systematic effect in the measurements (cross-talk effect). This aspect hardly finds literature attention, but it can play a major role if the safety of the structure is monitored with wire potentiometers. This paper focuses on the estimation and compensation of this cross-talk effect, to provide reliable measurements in the monitoring of slender structures. Numerical and experimental validation of the proposed approach demonstrates the need to compensate for the cross-talk effects for a proper uncertainty estimation and then the validity of the proposed method.