A Methodology for the Clusterisation of Communication Towers on the Basis of Their Structural Properties and Loads

Optimising the O&M activities related to the proprietary infrastructure assets is crucial for a company successful management. Nowadays this topic is gaining even major hype due to the accessibility of new technologies such as big data. Such reasons push the industrial market towards a pursue of effectiveness and efficiency of maintenance strategies. This work inserts itself in this context: born from a collaboration with a company involved in the sector of telecommunications, it aims at the determination of an heuristic model able to estimate, as far as their towers are concerned, the probability of a structural failure. The initial phase is about the analysis of the company’s database to identify the variables to be considered in the study. Then the focus is shifted on the research of a synthetic index that could summarise the contribution of the aforementioned parameters. Subsequently, a sensitivity analysis aimed at testing the robustness of such index is carried out, followed by a subdivision of the entirety of the assets into homogeneous clusters in terms of loads and structural properties. The clusterization is intended to feed the “risk” axis of a risk-impact matrix to be used to tailor the O&M best practices to apply to the fleet of communication towers. The predictive capability of such representation will be assessed performing physical checks on a number of towers sampled from the matrix.