Abstract:
Over the last few decades number of tall buildings has seen an exponential increase. At present tallest ever building under construction measures 1000m vertically, and the height of future buildings is likely to be even higher. Monitoring the behaviour of tall buildings using them as living laboratory is extraordinarily important in an international context to optimize its performance. Structural health monitoring is a new paradigm which facilitates the purpose of monitoring buildings or any other infrastructure on real time basis. Structural health monitoring has seen various advanced developments in recent past. Wired sensor networks were used to monitor the target at the beginning. In modern days wireless network with higher number of nodes is used to monitor the target very precisely.
In this research a wireless sensor network which is capable of sensing ambient vibrations in terms of accelerations was developed. The sensors mounted in each node are capable of measuring very small vibrations (1 mg range). The communication between each node was established using wireless network protocol, sub-1 GHz (Radio waves) which is very efficient in terms of long-range communications, power consumption and penetration through obstacles. When it comes to collecting data using wireless sensor networks, there are inherent challenges such as time synchronisation, scalability, packet handling (collision), packet loss, data storage, power consumption etc. However, accuracy of time synchronisation was identified to be the most crucial problem as far as interpretation of results is concerned. Two methods of time synchronization were checked in laboratory level. One method is known as receiver to receiver model and other one is centre to receiver model. Wireless sensor network was checked for its performance in laboratory level and accelerometers were calibrated using shaker table which has accurate analog sensors and digital accelerometers which are already calibrated in laboratory level.
Completely developed wireless sensor network which is capable of collecting synchronous data, was established in a target building of 48 floors with 185m height. The locations of sensors were predetermined using mathematical model made using finite element package, ETABS. Bi directional acceleration data was collected with sampling frequency of 100 Hz. Collected data was chunked and converted into frequency domain from time domain using fast Fourier transform algorithm and modal damping ratios, peak acceleration corresponding to particular frequency and modal displacement were extracted.
Extracted modal damping ratios were compared with the damping ratios suggested in various codes and it could be observed that the calculated modal damping ratios are higher than the
iii
values that are suggested by various codes. The mode shapes plotted using the building response data showed a good agreement with the mode shapes produced by Operational Modal Analysis.
