Prediction of non-woven geotextiles' reduction factors for damage caused by the drop of backfill materials

The need for sustainable solutions in geotechnical works has encouraged the investigation of recycled con-struction and demolition wastes (RCDW) as backfill material. The possible damages caused by the launching (dropping) process of this "new" backfill material (RCWD) must be quantified for its combined use with geo-synthetics. This study evaluated the influence of the backfill's grain-size distribution and the geotextile's mass per unit area in the damages caused by the launch of RCDW material and aimed to provide a prediction equation of the reduction factors. Five RCDW materials were launched from 1.0-m and 2.0-m height over five non-woven polyester needle-punched geotextile and specimens were exhumed to be tested under tensile. Databases were created with the results and subjected to machine learning to obtain a prediction equation for the reduction factor's values. The results show that the damages caused by the dropping height are complex. The 1.0-m in-crease in the drop height and the increase in the geotextile's mass per unit area cannot be associated with an increase in the damage. The geotextiles were more affected by the backfills with uniform gradation. A reduction factor's prediction equation is presented considering the three variables investigated (geotextile, drop height and backfill material classification). The artificial neural network is a more interesting solution than multiple linear regression since it does not possess several application criteria and provides more accurate predictions.