Advanced maintenance strategies for infrastructure assets such as bridges or off shore wind turbines require actual and reliable information of the maintenance status. Structural health monitoring based on vibration sensing can help in supplying the input needed for structural health monitoring applications. However, large structures require a sensor network with a large amount of sensors for vibration monitoring. Approaches with wired sensors and a central processing unit can raise scaling problems with respect to data processing and cabling. Wireless vibration sensors solve the cabling issues but due to limited communication bandwidth not all the data can be communicated, so the data processing has to be decentralized. Our goal is to develop a wireless sensor network which is capable to monitor the development of the eigenfrequencies and mode shapes over time using a decentralized approach. In this paper this network approach will be addressed and strategies for Operational Modal Analysis and decentralization aspects are examined. An experimental setup is available for performing experiments. This setup is a scale model of a wind turbine structure, which can be modified by adding masses at different locations. Experiments are performed with wired accelerometers to supply a dataset which can be used to study various approaches for distributed processing. In parallel, low cost wireless vibration sensors are being developed. The performance of these sensors will also be discussed.