To investigate self-replenishing on surface-structured composite coatings a dual simulation-experimental approach is employed to study the decisive role of polymer-air and polymer-particle interfaces. Experimentally, the composite system consists of a cross-linked polymer network with fluorinated-dangling chains, embedding colloidal SiO2 nanoparticles which are incorporated in the network via covalent bonding. These particles provide the desired surface structure at the air-interface before and after damage. Any damage replicates the rough surface, while the polymer layer on top of the particles serves as source of low surface energy groups which are able to reorient towards the new air-interfaces. Using coarse-grained simulations details of these self-replenishing composite systems are revealed such as the minimum thickness of the polymer layer necessary for providing optimal self-replenishing ability and the distribution profile of the dangling chains at the various interfaces. The principles and dual approach reported here may be applied to other self-healing composite systems with applications in self-cleaning, anti-fouling or low adhesion materials.