Effectively detecting and localising structural damage is crucial for ensuring the safety and longevity of engineering structures, making Structural Health Monitoring (SHM) an essential field of study. Vibration-based methods are often used to identify and localise occurring damage. These methods are based on monitoring the dynamic behaviour using recorded acceleration or strain data and detecting changes in this behaviour. For a more precise localisation, the observed changes are reproduced in a simulation model by means of optimisation to obtain more meaningful information. These model-based approaches are often applied in the context of digital twins. However, the position of the sensors used to record the applied data greatly impacts the localisation results. Several approaches for optimising sensor placements are available, yet their impact on the damage localising capabilities of model-based approaches is rarely analysed. This contribution uses a stochastic multi-objective optimisation approach to take parametric uncertainties into account in the sensor placement optimisation. Further, the effect of the optimised sensor layouts on the localisation results of model updating is investigated. The approach is applied to the Leibniz University structure for Monitoring (LUMO), a 9m tall lattice tower with several reversible damage mechanisms. Modal properties are identified for different sensor setups with varying numbers of sensors. A stochastic multi-objective model updating procedure is applied using the identified modal properties to localise a damage scenarios in the benchmark structure under examination. The results of the model updating procedure are compared for different optimal sensor setups, enhancing the insight into the impact of the sensor setup itself. The findings highlight the importance of strategic sensor placement in SHM applications, enabling more reliable model-based damage localisation.