PSI - Issue 17

Pavel Steinbauer et al. / Procedia Structural Integrity 17 (2019) 799–805 Author name / Structural Integrity Procedia 00 (2019) 000 – 000

802

4

The modal damping is difficult to be used for construction change assessment. Determination of modal damping is erroneous and usually covered by ambient influences (e.g. ground, sand damping and their changes due temperature and humidity changes). Frequency response function measured at suitable point of the pole (outside nodes of the structure) is sufficient to determine pole damage. The conclusions can only be made using repeated measurements of pole’s FRFs. Eigen frequency differences between similar, healthy poles are significant. Unfortunately, realization of experimental modal analyses measurement cannot be done by unskilled worker even for simple cases. It is also even more time demanding then pole loading by static force The paper introduces pole health detection method, based on eigen frequency shift determination using continuous acceleration measurement of the pole motion using only ambient excitation (mostly coming from the wind or traffic). The measured acceleration data are evaluated directly by local processor. The processor is integrated with sensitive MEMS accelerometer into one unit installed on the top of the pole. The measurements are carried out in selected intervals during whole pole lifetime. In the beginning, the system is initialized and trained. The markers of particular pole are stored. Later on, actual markers are regularly evaluated and compared to initial markers. Thus the methods is independent of the pole type, does not need excitation, special calibration or installation. 4. Device Concept

Fig. 4 Measurement system structure a and data flow b The concept is integrated into smart sensor (Fig. 4). The sensor is equipped with sensitive MEMS accelerometer. It carries out long term acceleration time series measurement and performs frequency spectrum calculation using Fast Fourier Transform (FFT). Data are evaluated based on eigen frequency shift and other quantities. The long time series enables to determine frequency spectra with very small frequency step. FFT is computationally demanding, but can be performed in longer time periods, so powerful processor is not needed. The change of pole markers is aggregated into 8bit output key value, which reports state of the smart sensor’s algorithm itself and the pole’s state change. The smart sensor is connected via wireless network grid (IoT Sigfox, IQRF etc.) to the central monitoring server or cloud. The pole regularly reports its state to central server. Inspection and replacement can be then aimed at problematic areas only. The sensor grid reports are also evaluated and compared on central level, so the areas with higher amounts of change reports can be inspected personally. The smart pole replacement based on predictive maintenance principles is more appropriate and enables to achieve both higher safety and significantly decrease maintenance costs. However, there is huge number of poles to be treated (about 1 mil. poles in Czech Republic only). .