SoLid is a reactor neutrino experiment, located at BR2@SCK-CEN (Mol, Belgium), searching for very-short baseline neutrino oscillation (<10 m). Its main objectives are to confirm the so-called Reactor Antineutrino Anomaly (RAA), and to test in-fine the existence of light sterile neutrino(s). This experiment is based on a new kind of neutrino detector specially designed against background. The fiducial volume is made of composite solid scintillators (polyvinyl-toluene or PVT and $^6$Li screens), which is compact and highly segmented (8000 voxels/m$^3$). As most of reactor neutrino experiments, neutrinos are detected through Inverse Beta Decay reaction (IBD), which produces a positron and a neutron. To perform a sensitive oscillation analysis, one of the most important requirement is to control perfectly the detection efficiency, which is in IBD case directly driven by the neutron detection efficiency. Regarding the SoLid setup, the challenge arise from the very large number of cells (12800). This proceeding will present the calibration strategy adopted by the collaboration to determine the neutron efficiency at percent level, by the use of two automated systems : CALIPSO and CROSS. Those last were designed in order to irradiate each 12800 cells, with well-calibrated neutron sources (AmBe and $^{252}$Cf). This proceedings also present the Monte-Carlo (GEANT4) simulation used on one hand to estimate the geometrical efficiency effects, and on the other hand to study the systematic errors arising from the neutron transport calculation.