The basement membrane (BM) is a nanoporous extracellular matrix that surrounds most tissues and blocks the passage of cells, primarily composed of covalently cross-linked collagen IV fibres and laminins. While BM breaching has traditionally been thought to be mediated by protease-mediated degradation, the failure of protease-targeting clinical trials to reduce metastasis suggests the existence of alternative protease-independent mechanisms. Recent studies indicate that invasive cells extend filopodia capable of remodelling plastic extracellular matrices. However, the covalent cross-links in collagen IV fibres are very strong, prompting the question of how filopodia might facilitate BM invasion. Collagen IV fibres undergo turnover---a dynamic process of protein renewal that may create transient weak spots in the BM. We hypothesise that filopodia exploit these weak spots during turnover to initiate and expand pores, enabling protease-independent invasion. We propose a mathematical biophysical model to test the plausibility of this mechanism using biologically relevant protruding and turnover conditions obtained from experimental observations in the literature. Invasive cells are represented as energetic biomembranes using geometric-surface partial differential equations, allowing the formation of filopodial protrusions, while the BM is modelled as a barrier with collagen IV cross-links that stochastically transition between active and inactive states. The results of the model contrasted with experimental observations identify two subpopulations of filopodia in invasive cells: thin, short-lived filopodia that contribute to global BM degradation, and long-lived, widening filopodia that locally stabilise and enlarge pores where invasion can eventually occur. Under suitable conditions, the model predicts that random turnover and filopodia can synchronise, leading to progressive pore enlargement. Further, pore enlargement arise from the collaboration of several filopodia entering and leaving the same region of the BM at different times. Although our results cannot demonstrate that this mechanism occurs in vivo, they place turnover as a plausible contributor to protease-independent invasion.