Strategy for required data reduction in the practical implementation in a low-cost electronic platform of an index for damage assessment of seismic dampers

This paper presents a strategy to reduce the number of data required for the in-situ health evaluation of hysteretic energy dampers used for seismic protection of building structures. Such an optimization is essential in order to implement in practice the numerous indices now available for damage assessment of structures in real time, in the context of Structural Health Monitoring (SHM). In general, damage indices are implemented and verified in lab systems entailing expensive and very specialized equipment, which permits the use and management of large amounts of data. For real-time field applications, low-cost embedded computers with limited calculation and data recording capacity may be used; however, they require thorough study of the influence of variables involved in the algorithms for calculation of the damage indices, with ulterior optimization and required data reduction. While the proposed strategy can be followed in broader applications and SHM techniques, the present paper focuses on its application to data coming from vibration tests carried out to evaluate damage of a particular type of hysteretic damper, the Web Plastifying Damper (WPD). The WPD was patented by the University of Granada for the passive control of structures subjected to earthquakes, through a non-parametric damage index previously developed by the authors. This study describes two steps for reducing data: (1) An in-depth study of the calculation time spent on each step of the theoretical algorithm; (2) A parametric study of the influence of key signal parameters-window length, number of windows and duration of the signal-on the damage index calculation in order to minimize the resources needed. Results show a successful optimization that permits the damage index to be calculated with the low-cost Picocom platform.