Cross-correlation methods for enhanced monitoring and health assessment of wooden poles

Despite the availability of glass-fibre materials, wooden poles are still largely used by utilities due to their easy-of-use, versatility and cost. Poles mechanical functionality and structural integrity should be carefully monitored to avoid breakdowns or failures. Economical and environmental considerations recommend the adoption of non-destructive assessment techniques. In this context, a promising approach relies on the analysis of stress waveforms induced by pre-defined vibrational stimuli, e.g. hitting the pole surface with an instrumented hammer. In particular, the basic idea is to represent the vibrational data as a linear combination of three damped sinusoidal components and infer the pole health condition from the stress wave spectral content. In fact, wooden poles responses account for interesting mechanical features, such as stiffness and flexural rigidity. Unfortunately, the responses are also influenced by a large variety of external factors, such as ground foundation solidity, over-head cable connections with neighbouring poles or other infrastructures, and eventual external surface treatment. This paper presents three promising methods which exploit a proper formulation of the measurement cross-correlation matrix to drastically reduce the noise level and significantly enhance the probability to correctly discriminate the informative components from the non-informative ones. Both synthetic and experimental data are employed to characterize the methods performances in terms of success probability and estimation accuracy.