Investigating Performance of High-Rate GNSS-PPP and PPP-AR for Structural Health Monitoring: Dynamic Tests on Shake Table

Abstract

This paper investigates the usability of Global Navigation Satellite System (GNSS) Precise Point Positioning (PPP) methods, traditional PPP with a float-ambiguity solution and with ambiguity resolution (PPP-AR), in structural health monitoring applications based on experimental tests using a single-axis shake table. To evaluate the performance of the PPP methodologies, harmonic oscillations of the motion table with amplitudes ranging from 5 to 10 mm and frequency between 0.1 and 3 Hz were generated representing a wide range of possible structural motions. In addition, ground motion similar to those experienced during a real earthquake, the 1995 Kobe earthquake, and step motions were generated on the shake table. GNSS PPP–derived positioning results at 20 Hz were compared, in both of the frequency and time domains, with reference data comprising LVDT data and relative positioning data. Results show that both PPP methods’ measurements can be used in the computation of harmonic oscillation frequencies compared to the LVDT and relative positioning values. The observed amplitudes of the harmonic oscillations are slightly different from the LVDT values on the order of millimeters. The results of a step motion experiment demonstrated that PPP-AR is better than traditional PPP in exhibiting quasi-static or static displacement. Moreover, the capabilities of traditional PPP and PPP-AR methods are evaluated with respect to the natural frequency of a small-scale structural model excited on the shake table. The frequency spectrum of this small-scale structural model derived from the PPP methods is consistent with finite-element model (FEM)–predicted values and relative positioning. The research presented here demonstrates the potential of the high-rate GNSS PPP and PPP-AR methods to reliably monitor structural and earthquake-induced vibration frequencies and amplitudes for both structural and seismological applications. Specifically, all results reveal that high-rate PPP-AR is more accurate than traditional PPP for both dynamic and static displacement detection.